Effective Sales People
Commit to Providing Value

Brendan McAdams invited me to take part in his “Let’s Chat Health Tech” podcast and we talked about the importance of discovery to uncover customer needs, why salespeople who commit to providing value are the most effective, and the importance of playing a long game. Bredan’s summary of key points:

• Shifting from closing transactions to creating enduring customer value
• The role of discovery and fit in complex sales processes
• How to build trust and credibility within customer organizations
• Using project management frameworks for disruptive technologies
• The importance of collaboration and co-creation during proposal development
• Strategies for managing expectations and avoiding surprises
• Recognizing and supporting the shift from exploration to operationalization
• Early-stage sales tactics for managing technical requirements and customer objections
• Navigating complex decision-making groups and internal gatekeepers
• Maintaining a long-term perspective beyond initial deal closure

Edited Transcript

Brendan McAdams: Hello, it’s another episode of Let’s Chat Health Tech. this is a little less health tech and a little more just Let’s Chat sales. But we have the good pleasure of having one of my good friends, Sean Murphy. Sean is the CEO of SKMurphy Inc. He has been an entrepreneur and intrapreneur and serves as an advisor to a number of startups. He helps them explore risk-reducing business options and building scalable, repeatable sales process.

So our practices overlap quite a lot. He focuses on early customers and early revenue for startups, especially software companies and helps engineers. Don’t you have a degree from MIT or something like that?

Sean Murphy: Not MIT, no, the MIT of the west, Stanford. For your listeners on the East Coast: it’s a small college out west on the other coast.

Brendan: Stanford is legit.  You take a wonkier approach to sales than I do, drawing from a diverse set of resources which is based relationship. My approach is more intuitive and based on trial and error.

For today we’ve decided to talk about complex sales, selling to enterprise buyers, and the importance of committing to  delivering sustainable value.

Sean: I think that’s one of the hallmarks of your approach that I’ve always admired. Too much sales advice is focused on controlling the outcome, closing the transaction, and getting the purchase order. But the reality is, in the long run, if it’s not a good deal for the customer then it won’t be a good deal for you.

Brendan: I totally agree. There’s a natural flow to a good deal that takes a while to learn. But you experience it over a period of time, and you start to see the deals that are likely to close, are likely to be successful, and will deliver long-term value for both the vendor and the customer. We can spend time breaking down what the ingredients of a good deal are and how to manifest them more often.

Put Customers On a Path That Takes Them Higher

Sean: So my context: I work with early-stage teams, typically two to five engineers or scientists who have a complex technology, and they’re looking for early deals. So they’re trying to figure out exactly what they’ve got. Often, when you’re selling something that provides a significant boost in productivity, cost savings, or cycle time, the customer needs to change their process. It’s not always a form, fit, and function insertion.

I advise clients to treat this like a project management challenge: they are scoping and managing a project to install capabilities that deliver business value to the customer. I think a project management framework is not a bad framing, at least for more disruptive or discontinuous technologies.

Brendan: That is an interesting way to put it because I coach clients to view themselves more as consultants than sales people. They should guide customers to a successful outcome, in much the same way a project manager would.

Sean: I agree. The first step to delivering value is making sure that you understand the customer’s problem or need. Then you have to make sure there’s a fit between that need and the capabilities your product offers. Sometimes entrepreneurs skip that first step and all the discovery that goes with it, assume their product is a good fit, describe a few features, and get started on closing the deal.

Brendan If you start by trying to understand the customer’s situation they are going to feel less rushed. They will appreciate that you are trying to examine all the things that need to happen and might go wrong. I think that gives the customer a much more kind of comfortable feeling about moving forward.

Sean: I work primarily with bootstrappers who are often trying to close deals to meet payroll, so to speak, or keep the lights on. But even in that context, I think the reality is that the urgency comes only from the customer’s need. No matter how much you need the deal, if you can’t connect it to objectives, timelines, and results the customer needs, you can’t inject urgency.

Brendan Right. We all want to have things happen more quickly. A salesperson is often working against time–end of the month, end of the quarter, or end of the year. So delays work against you. The more you let things go, the more things could happen in the interim. The way you compress time is to be more responsive, to get things done sooner on behalf of the customer. Rushing the customer often slows things down.

Trust Makes Things Smooth, and Smooth is Fast

Sean: I think the kind of deals you’re involved in, and many of the deals I’m involved in, there may be an internal champion and one key decision maker, but there’s often a group or a committee or a set of people that have to sign off in different ways or at least contribute to and shape the decision. There’s a Navy SEAL mantra that “slow is smooth and smooth is fast.” If you lay the groundwork by building relationships, in the long run, that actually speeds things up.

Brendan As the customer recognizes that you’re genuinely interested in doing the right thing on their behalf, you establish shared trust, and the amount of due diligence that they feel the need to do goes down. It doesn’t go away, but goes down. The reverse is also true: if they start to believe you are only focused on the deal, then they feel they need to double-check everything.

Some of the ways you can look out for them with your project manager hat on are to identify what’s going to get in the way of a pilot or initial project delivering value. What might cause it to fail? Can you address them now? If you can’t, it may be a case where “no” is the right decision. If you can’t deliver value to the customer in a meaningful and measurable way, you’re probably better off not selling that deal.

Because if you’ve got a customer who doesn’t succeed, that’s a big black eye for you with other customers, support issues, collections, and more. So if you’re really taking the long view and playing the long game, you as a salesperson need to see that deal succeed. Otherwise, it’s an albatross hanging around your neck.

Sean: That’s a very good point that often gets overlooked in the heat of the moment. There is normally one community that you and the customer are members of, and often several. It may be organized around a particular technology,  discipline, profession, or geography. You will develop a reputation within multiple communities that the customer evaluation team members are part of.

We have a simple sales model. The customer has to understand your product, believe it can do what you promise, and find a reason to act. That second step, the belief step, typically relies on the trust you have established in the way that you have engaged with them and the credibility you have established with people in communities the customer is a member of and can verify.

Brendan I think that’s absolutely true. And I believe that over time, if you’re in sales, and frankly, if you’re a founder of a company or high up in a company, this applies just as well. And there’s certainly a certain component of your expertise in sales, regardless of whether or not you’re the CFO or COO or the CEO or actually in charge of sales.

Over time, you become more and more recognized for your reputation and expertise, which supports your idea of acting as a project manager and making things happen. I hope you don’t think I’m overworking that metaphor.

You Must Talk to Enough People to Get a Holistic Perspective

Sean: I don’t. Another aspect of the project management paradigm is the need for joint scoping of needs by talking to multiple people in the customer organization to get a holistic perspective on the total set of requirements. Sometimes you can disqualify yourself when it becomes clear there is a misfit between their needs and your product’s capabilities. To your earlier point, it’s better to detect the misfit before you’ve taken their money than after you’ve taken their money and can’t deliver, and now everybody’s unhappy.

Brendan: You made a very valuable point: you must get around inside that organization. You want to talk to as many people as possible to get that holistic view. That’s going to allow you to communicate with more people more effectively in the language they understand and that resonates with them. You’ll be able to use their examples, their terminology, and their metrics. And by doing so, you’ll make it much easier on yourself and your champion inside the organization to get this deal across the finish line.

Sean: Many times, especially early in the life cycle of a new technology or a new problem space, different terminology is used in different companies or regions. Different terms may actually mean the same thing. Figuring out how they talk about the problem and adapting to their language is important.

Another aspect of soliciting input comes when they ask you for a proposal. It’s better that you work on it together at the whiteboard or in a shared document where you are both actively engaged and collaborating on a shared plan for moving forward. Don’t pour your notes into ChatGPT and ask it to write a 12-page proposal. These don’t get read, and often sending it to a prospect can end the conversation–you never hear from them again. The best proposals and project plans are co-created with key members of the customer organization.

Brendan: Going back to this project manager example, part of your job as a salesperson is to create shared recognition of the problem you are trying to solve. We agree that this is what we are trying to achieve together. And there may be a fair amount of difficulty and discomfort in the coordination and politicking to get that done. You, your champions, and the other people in the organization need to figure out and agree on the steps, moving parts, and obstacles to be overcome.

You also need to bring in others from your company to review and commit to the proposed solution. If you’ve done your job well, it’s not two sides but everyone on one team looking at the challenges and the plan together. Done right, you can hit their target and achieve all of the business objectives your firm needs to make this a good deal for both parties.

Negotiation is a Means to Arrive at a Solution Jointly Agreed To

Sean: That makes sense. Project managers negotiate; salespeople negotiate; negotiation is a means to arrive at a solution jointly agreed to. At the end of the day, if the customer doesn’t get value out of what your product can do for them, then nobody wins. Maybe you get a one-year deal, but as you pointed out, when it doesn’t provide value, it leaves a lasting bad taste in everybody’s mouth. So you have to get multiple perspectives, because no one person understands all of the complexity involved in a successful transition.

In particular, the champion may understand what’s needed for a successful proof of concept or pilot project, but may know less about what’s needed to scale up from there to operationalize the solution fully. What procedures do we have to change? What’s going to change about how the organization does business to take advantage of this product? I spend a lot of time helping my clients help their customers scale up from an initial success.

If you are selling to a freelance website developer, that person is the business decision maker, the manager, and the user. So you can get a decision very quickly, and they can figure out what it will take to use it. There is no scaling beyond making that one freelancer successful.

But it becomes complex when you’re trying to support a team, department, or larger group spanning organizational boundaries. It becomes much more complex when your product affects how the customer does business with their customers or suppliers. As the scope of the change increases, there is more negotiation and often more re-engineering of existing procedures and workflows. All of which has to be mapped out to make the first implementation a success.

Brendan: Another benefit that flows from this consultative project management approach that involves the customer in the design of the solution, mapping all of the moving parts, and identifying evangelists is that you are in a much better position to guarantee success. You can plug your team members into the right conversations and prepare for various eventualities you may need to manage.

For example, here is the person who will be your account manager. This minimizes surprise, and large organizations don’t like surprise. They want to know that this is going to happen, and at this point, this is going to happen, so that they can anticipate and plan ahead. It works well in setting expectations in both directions: we’re going to need this from this department, and we’re going to need these people to be involved in this training.

Of course, if people know about these things ahead of time, they can object, and you may have a potential adversary. But doing it early means it’s not a surprise, you have more time to make adjustments, and they are much more inclined to participate enthusiastically because they can plan for it.

Sean: I am OK with tough questions, they are normally an indication you are speaking to a serious prospect. I would much rather be talking to someone who ask, “How do we do this? How do we do that? How do we operationalize this for our group?” The people who a just focused on getting a quote have already slotted you into third or fourth place as a makeweight. Large companies often require three bids and they invite you to bid so they can say, “we’ve evaluated three options and it turns out that the guy we like when we started turned out to be best.”

Tough Questions All the Way Around Are a Good Thing

Brendan: I would argue that the converse is also true. When you are in a sales or consulting role, you are an expert knowledgeable about your side of the solution, and you’re certainly learning more and more about their side of the problem. Every customer has their own environment, with unique components and aspects to their workflow. They may not face the same obstacles, but you will find common patterns, which will enable you to ask tough questions as well. Like, what could go wrong? Who will oppose this solution? Whose ox is gored by this decision? Any of those sorts of things. I mean, you’re entitled to know those things as that outsider who’s evaluating and consulting on their behalf. And so I think tough questions all the way around are a good thing: they are something you should not only expect but also encourage.

Sean: Another thing I’ve seen happen, especially early in a product’s life, is that you have a clear view of what the complete feature set should be. Then a customer says, “We’re going to need an interface to this software tool or this piece of equipment.” But because you are blessed with insight, you say, “Actually, we plan to support this other software and this other brand of equipment.”

When they respond, “OK, but we still need it. If you’re not going to do it, we’ll do it.” What you don’t realize is that they are giving you a new requirement that other companies like them, perhaps firms above a certain size or in a particular vertical, will ask you for. By turning over the implementation to them, you’ve forfeited the chance to work with someone to debug a more general solution, because they are not going to make that interface available to their competitors.

In B2B, if you’ve sold to 30 or 40 customers, you should have a pretty good sense of the market and where this particular prospect fits within the distribution of what others are selling and what most customers need. You have a better sense of whether their request is a one-off that interfaces with legacy software or equipment, or it’s on the roadmap and should be accelerated. If it’s the former, then that’s their problem or perhaps a professional services opportunity.

But early on, feature content and interoperability requirements are less well settled. And I think that argues as well for figuring out what it’s going to take to get this thing into real production use to be part of the critical infrastructure.

Brendan: I think that’s one of the reasons early-stage companies benefit from a consultative sales approach. This approach also benefits companies in a growth stage that face significant competition in the marketplace, where new features and functionality are expected. Having a salesperson who can act as a consultant to customer organizations will be invaluable because they can surface the requirements for what is needed in the next release or on the product roadmap.

The other thing a good consultative salesperson can do is manage customer expectations: “Here is why you can’t have that feature yet. That’s going to take three months or six months, or we need six other customers to buy into it to make sure that it is actually functionally required in the marketplace.” They may also say, “That’s a one-off, we’re going to need $50,000 for development fees to bring that in and to support it.” That’s what I mean by managing expectations.

Sean: We don’t do a lot of work on the medical side, but one related area is medical equipment that is sold for “research use only(RUO),” which means selling to research labs and what are called central labs. These buyers have different requirements from what is needed to get deployed in a hospital or a clinic, and represent a much smaller market niche, but it’s a way to get started.

A second area is to attack the veterinary market, because it’s easier in many ways. Over time, you can migrate to clinical use for humans.

In both cases, you have to satisfy that initial community, but you’ve got to be alert to the fact that you don’t really have the full product spec for what it’s going to take to get to the major market or the major opportunity. On the other hand, if you don’t make the researchers or veterinarians happy, you won’t get beyond them to the major market.

Brendan: The corollary for me is that I’ve coached people in certain health tech companies that need FDA approval. In some cases, you can work around that in the short term to achieve some adoption that provides feedback and outcomes, generating data for analysis and case studies. One of the best ways to do this is to work with DoD: if your solution can assist in a military setting, they don’t have to wait for FDA approval if the results are compelling enough. They can drop it into a location, start testing it out, and prove it out.

Then you can start the approval process with a fair amount of traction. That out-of-the-box approach can sometimes move things forward quickly. Good salespeople know how to think outside the box, know how to consult, and know how to examine things from a lot of different angles

Sean Murphy: I think that’s a good example of looking for a win-win. You must have some proof of efficacy for a problem that the DOD is facing, right? You’re not going there just because they have a more relaxed standard. In some ways, they have much more stringent standards. They just don’t need to wait for the FDA.

The Sale is Never Really Over

Brendan: Right.

So let’s start to wrap this up. We’ve talked about the need for a good discovery process to uncover customer needs and make an offer that provides compelling long-term value. I think there are two implications: the sale is never really over, and there is a lot to learn from a customer saying no.

The idea of “closing” someone has always made me feel uncomfortable. The customer is never permanently closed. You can have a signed contract and full payment, but that doesn’t mean anything is really over. In many ways, the real work of supporting the customer has only begun. And before that, until they have seen results that demonstrate your value convincingly, there is always the possibility that the deal is not done.

You need to make sure that the customer has the opportunity to say no. It is very important to let the customer examine the situation and accept when they say, “This isn’t a fit for me.” If you make them feel comfortable doing that, the less likely they are to look for excuses to say no, the more likely they are to try to figure out why they should do something.

I’ll stop rambling on here in just a second, but the thing I’m thinking of is that many customers, especially in healthcare, don’t want to do something else. There are already too many things they must do. So you do have to address that. But at the same time, if they’re really going to say no, right, because they just don’t have the bandwidth, they don’t want to do it, they’re lazy, or they’re just a hidebound old school organization, then you want to know that. And you want to know that soon rather than later. You don’t want to go through six months of evaluation and a dozen 30-minute sales calls that ultimately go nowhere.

Sean Murphy: There’s a lot to unpack in what we covered: I agree with your conclusions. To rewind to an earlier point about handoffs: in a small startup, there is a lot of overlap between sales and operations. As the business grows, there’s a much greater need for a handoff between people focused on new opportunities and those handling day-to-day customer interactions.

I think over the course of a deal, the customer also shifts from an exploration mode to a delivery or operations mode. OK, now we’ve decided to move forward and need to operationalize the capabilities the product offers. Operations people are usually less involved in the early stages of an evaluation. I think when you are open with the customer up front, you’re going to get more of the real reasons for a no. And you’re going to get less “performative” testing and more verification focused on what they believe are their real needs.

If you do three demos and learn something from each, to me, that’s winning. Even if the sale doesn’t close, I place a high value on learning with a new product. The best case is that you learn a lot and close the sale, but if you don’t know why you closed the deal, that’s not good, because you were just lucky and aren’t building a repeatable, scalable process. And the real reasons for a no are very valuable because it’s definitely not good when you lose the deal and are mystified, because you are going to make the same mistake again on other deals, and you might chase opportunities you shouldn’t.

In the same way that the customer shifts from discovery to delivery, the sales team shifts from explore to exploit as they learn what works. They still do discovery, but they are executing a debugged process instead of debugging their process. The first purchase order or initial sale is just one milestone in the evolution of an ongoing business relationship.

I liked your points. I just wanted to add a few more items we covered.

An Enterprise or Complex Sale is a Long Game

Brendan McAdams I think that’s accurate, certainly for an enterprise or a complex sale, where the customer is not buying a commodity but a complex product. They are buying a result that won’t happen right away. It will require education, implementation support, and configuration to their specific needs. That’s a long game. You’re playing the long game.

So for our listeners, if you’ve got questions about this or have things you want to add, please leave a comment or reach out to us.

Related Blog Posts

• Early Sales Efforts Foster Value Co-Creation
• Entrepreneurial Mindset: Create Value For Others
• Price Based On Your Value To The Customer’s Situation
• Successful Bootstrappers Are Trustworthy Salespeople Committed to Customer Satisfaction
• Price, Value, and Your Prospect’s Perception of Risk
• The Robin and the Cherries: Sharing Value You Co-Create With Customers
• Newsletter: Crafting and Testing Your Value Hypothesis
• Building a Business Requires Building Trust
• I Heard You Went to Stanford

Brendan McAdams

I have tremendous respect for Brendan’s insights and expertise. I have done two Q&A’s and a joint article as well as inviting to present twice at Lean Culture. I also reviewed his book “Sales Craft” favorably.

• Drawing a Map of the Customer Organization (Q&A)
• Cultivating Luck in Business Endeavors and Relationships  (Q&A)
• Why is an Enterprise Sale So Complex? (Joint Article)
• Brendan McAdams on 5 Sales Fundamentals for Startup Founders  (Presentation)
• Navigating the Startup Sales Maze: Brendan McAdams’ Expert Advice  (Presentation)
• Brendan McAdams on Sales Craft: Proven Tips to Elevate Your Sales (Book Review)

A Postscript on “Makeweight”

My father was a maritime who specialized in the inland waterways of the US. One term he would use from time to time was “makeweight,” which was low value cargo loaded on a barge to allow it to be steered more easily. It meant something of no consequence, similar to packing peanuts or other filler materials. In “Map Your Customer’s Buying Process Before Sending a Proposal” I blogged about a risk startups face when selling to enterprise buyers:

You May Be Column Fodder: More often than not you are actually “column fodder” or a makeweight needed so that they can prove to their boss or the purchasing/finance team that they did a thorough job and solicited three bids. Especially if you don’t know much about their situation and they have not asked for a detailed demo you need to proceed a little more slowly.

Column fodder is a back formation from “cannon fodder,” a military term from the 19th century for troops of low quality who could be fed to the enemy’s cannons to distract them from firing on higher quality troops. Another term I find appropriate for startups that find themselves in column 3 (and effectively out of consideration) is “dead rat on a stick.” The buyer uses the startup’s functionality and pricing to prod the incumbent into adding new features and/or cutting the price to match the startup’s offer. This is a tricky situation because they are encouraging all the way through the process to keep the startup involved to extract concessions from the other two vendors, but they have no intent to buy from the startup.

I have also blogged about “column fodder” in “Set Your PowerPoints on Stun” and “Early Proposals: Avoid Consulting For Free.”

Image Credit Trail Goes Up (c) Kevin Murphy, used with permission. I picked this because it looks rough but it’s actually paved with stones and is much more than just a cut trail. This is the level of rough and rough ready than most complex sales achieve. It’s not a paved road but it’s serviceable year round.

Quotes for Entrepreneurs Curated in May 2026

I curate these quotes for entrepreneurs from a variety of sources and tweet them on @skmurphy about once a day where you can get them hot off the mojo wire. At the end of each month I curate them in a blog post that adds commentary and may contain a longer passage from the same source for context.

My theme for this month’s “Quotes for Entrepreneurs” is theory and practice: their distinct value, differences, and how each informs the other.

+ + +

“In theory there is no difference between theory and practice, while in practice there is.”
Benjamin Brewster

I originally curated this in July 2019 and decided to use it as a theme for this month’s quotes.

+ + +

“One important idea is that science is a means whereby learning is achieved, not by mere theoretical speculation on the one hand, nor by the undirected accumulation of practical facts on the other, but rather by a motivated iteration between theory and practice.”

George E. P. Box in “Science and Statistics” (1976)

+ + +

“Practice is the exercise of an art or the application of a science. Theory is knowledge. Each presupposes and is dependent on the other. Theory is dependent on practice; it is a generalization of the principles on which practice proceeds. But there is no practice without theory: a man of practice must always know something, however little, of what he does and the outcome he intends.”

Sir William Hamilton, 9th Baronet (condensed from in Metaphysics of William Hamilton (1872) by Francis Bowen

Here is a longer excerpt for context (italics in original rendered as bold)

“The terms theory and theoretical are properly used in opposition to the terms practice and practical. […]

Practice is the exercise of an art, or the application of a science. […] It is not every one who is able to apply all he knows ; there being required, over and above knowledge, a certain dexterity and skill.

Theory, on the contrary, is mere knowledge or science.

There is a distinction, but no opposition, between theory and practice; each to a certain extent supposes the other. On the one hand theory is dependent on practice; practice must have preceded theory ; for theory being only a generalization of the principles on which practice proceeds, these must originally have been taken out of, or abstracted from, practice.

On the other hand, this is true only to a certain extent; for there is no practice without a theory. The man of practice must have always known something, however little, of what he did, of what he intended to do, and of the means by which his intention was to be carried into effect. He was, therefore, not wholly ignorant of the principles of his procedure ; he was a limited, he was, in some degree, an unconscious, theorist. As he proceeded, however, in his practice, and reflected on his performance, his theory acquired greater clearness and extension, so that he became at last distinctly conscious of what he did, and could give, to himself and others, an account of his procedure.

Sir William Hamilton, 9th Baronet (condensed from in Metaphysics of William Hamilton (1872) by Francis Bowen

This precedes the concept of “theory-in-use” developed by Donald Schon and Chris Argyris (See for example their “Theory in Practice“) “or the concept of “tacit knowledge” that informs our practice even when we don’t voice it directly.

I believe that effective teams share both a set of agreed-upon practices (at least at the level of working consensus) and shared mental models, mental maps, and theories-in-use. I have seen a lot of focus on leveraging cognitive task analysis to determine experts’ mental models and guide the training of novices and journeymen in a discipline. I think it may also be used to help teams reach a working consensus on shared mental models and shared practices.

In addition, customers always have a “theory-in-use” of why they are doing something or how a tool or process should work. They are not a blank slate waiting for you to bring them fire. The challenge is to ask the questions that unlock how they currently view their need or challenge so that you can anchor your explanation in their current understanding.

+ + +

“The great tragedy of science–the slaying of a beautiful hypothesis by an ugly fact.”
Thomas Huxley

Also applies to product hypotheses, better to uncover disconfirming facts pre-launch with conversations and low fidelity demos than post-launch.

+ + +

“The specific triumph of the technical imagination rested on the ability to dissociate lifting power from the arm and create a crane: to dissociate work from the action of men and animals and create the water-mill: to dissociate light from the combustion of wood and oil and create the electric lamp.”

Lewis Mumford in “Technics and Civilization” (1934)

The paragraph the immediately precedes this on make an interesting observation:

“Circular motion, one of the most useful and frequent attributes of a fully developed machine is, curiously, one of the least observable motions in nature: even the stars do not describe a circular course, and except for the rotifers, man himself, in occasional dances and handsprings, is the chief exponent of rotary motion.”

Lewis Mumford in “Technics and Civilization” (1934)

+ + +

“AI personalities and their owners should be subjected to fiduciary duty when they interact with users.

Fiduciary duty is a legal concept applying to an actor who is in a relationship of trust with another and whose position involves superior knowledge or skill. Classic examples include lawyers acting for clients, administrators of a legal trust acting for beneficiaries, executors for heirs, corporate directors for shareholders, and members of a partnership for one another.

A fiduciary relationship creates a duty to act in the beneficiary’s interest, to act in good faith, and to use reasonable skill and diligence. It includes—important in this context—a duty of confidentiality, with all information relating to the relationship kept confidential and private.”

Glenn Harlan Reynolds in “Don’t Stand So Close to ChatGPT” (May-3-2026) an opinion piece in the WSJ.

+ + +

“Space launch was a clear case where there was a large difference in efficiency between what was possible and what was done in practice before SpaceX. A large part of that was due to everything being locked in to what (just barely) already worked, with huge risk aversion. With national prestige or a half billion dollar geosync satellite on the line, speculative engineering ideas that might result in a public debacle were not welcome.

When failure is not an option, success can stay very expensive. You need to experiment to improve, and that fundamentally means being comfortable with failure. If you know it is going to work, it isn’t an experiment.

I have long believed that nuclear power today is in precisely the same state as space launch two decades ago, but the even more pressing question now is if semiconductor fabrication might also be.

On the one hand, Moore’s Law has been a sequence of heroic miracles of technology at the wafer fabrication level, grinding out hundreds of compounding small improvements.

On the other hand, fabs are “too big to fail”, and there are elements of extreme conservatism at play. Intel’s “Copy exactly!” fab development exemplifies that mindset – instead of every new building being an opportunity to explore and optimize processes, it was deemed more valuable to just replicate.

While each individual machine may be straining against physical limits of technology, it is possible that the systems orchestrating them all together could be far from optimal.

The explore / exploit axis is fundamental to all decision making, but human risk avoidance probably biases away from optimal exploration.”

John Carmack (@ID_AA_Carmack) May-7-2026 tweet

+ + +

“Without theory, practice is but routine born of habit. Theory alone can bring forth and develop the spirit of invention.”

Louis Pasteur

Inaugural Address as newly appointed Professor and Dean (Sep 1854) at the opening of the new Faculté des Sciences at Lille (7 Dec 1854). In René Vallery-Radot, The Life of Pasteur, translated by Mrs. R. L. Devonshire (1919).

+ + +

“Modern psychology teaches that experience is not merely the best teacher, but the only possible teacher. There is no war between theory and practice. The most valuable experience demands both, and the theory should supplement the practice and not precede it.”
Charles Kettering in Speech “Lap-Welding Students to Life” collected in “Prophet of Progress: Prophet of Progress: Selections from the Speeches of Charles F. Kettering” Edited by T. A. Boyd

+ + +

“Like body and soul, theory and practice are one, and like body and soul they are for the most part at loggerheads.”

Marie von Ebner-Eschenbach

+ + +

“Everyone talks about speed.
Faster execution, tighter loops, more opportunities, founders edging to their limits.

Push, push, push.

Pause. Breathe.

The speed at which founders move is critical to gain initial momentum. But it’s important to recognize that most founders end up in a marathon, not a sprint.

Saying this to founders who are already years deep in hardcore mode, anxious that they should still speed up.

In a marathon, persistence matters more than speed; discipline beats intensity, and smart energy planning to operate for 5-7 years becomes the ultimate moat.

Speed up when you have to, slow down when you can.”
Tom Ogrodzki  (LinkedIn)  May-9-2026 tweet 

+ + +

“In spite of illness, in spite even of the archenemy sorrow, one can remain alive long past the usual date of disintegration if one is unafraid of change, insatiable in intellectual curiosity, interested in big things, and happy in small ways. In the course of sorting and setting down of my memories I have learned that these advantages are usually independent of one’s merits, and that I probably owe my happy old age to the ancestor who accidentally endowed me with these qualities.”
Edith Wharton in “A Backward Glance” (1934)

+ + +

“An ounce of practice is generally worth more than a ton of theory.”

E. F. Schumacher in “Small Is Beautiful: A Study of Economics as if People Mattered”

+ + +

“Let me present an alternate future with new classes of creative people instead of new classes of obsolete people and an exponentially expanding sphere of creativity that people are proud of and motivated to participate in.”
Jaron Lanier in “Exponential Creativity” episode of Existential Hope Podcast [edited]

+ + +

“Nothing is more certain, than that improvement in human affairs is wholly the work of the uncontented characters; and, moreover, that it is much easier for an active mind to acquire the virtues of patience, than for a passive one to assume those of energy. […]

All intellectual superiority is the fruit of active effort. Enterprise, the desire to keep moving, to be trying and accomplishing new things for our own benefit or that of others, is the parent even of speculative, and much more of practical, talent. […]

“The test of real and vigorous thinking, the thinking that ascertains truths instead of dreaming dreams, is successful application to practice. Where that purpose does not exist, to give definiteness, precision, and an intelligible meaning to thought, it generates nothing better than mystical metaphysics.”

John Stuart Mill “Considerations on Rrepresentative Government” (1861)

+ + +

“Practice is the best of all instructors.”
Publius Syrus in “Moral Sayings“

+ + +

“The new unfair advantage isn’t bigger ambition. It’s narrower domain expertise nobody else can be bothered with.”
Ash Maurya

A narrow initial focus does not preclude growing a large firm over time, but it enables:

1. A better understanding of customer needs.
2. A higher level of customer intimacy.
3. A simpler product that’s easier and faster to develop.
4. Better word of mouth.
5. A profit sanctuary that larger competitors will ignore.
6. Establishing a beachhead in a niche market that has unique needs rewards honesty; it’s harder for larger firms to make promises they don’t keep and win on size instead of reputation.

I blogged about this in “Starting Small can Lead to Big Success.”

+ + +

“The world can doubtless never be well known by theory: practice is absolutely necessary; but surely it is of great use to a young man, before he sets out for that country, full of mazes, windings, and turnings, to have at least a general map of it, made by some experienced traveler.”

Philip Stanhope, in “Letters Written by the Earl of Chesterfield to His Son“

Quote is from Letter LXXVIII August 20, 1749.

+ + +

“While in theory randomness is an intrinsic property, in practice, randomness is incomplete information.”

Nassim Nicholas Taleb in “The Black Swan: The Impact of the Highly Improbable” (2007)

+ + +

“For the theory-practice iteration to work, the scientist must be, as it were, mentally ambidextrous; fascinated equally on the one hand by possible meanings, theories, and tentative models to be induced from data and the practical reality of the real world, and on the other with the factual implications deducible from tentative theories, models and hypotheses.”

George E. P. Box Science and Statistics (1976)

Entrepreneurs face a similar challenge: they need to make the effort to generalize from specific observations, stories, data points and other miscellaneous bits of evidence. But they cannot let their tentative hypotheses lead them to reject disconfirming evidence.

+ + +

“Instead of scientific research being set up merely at the instigation of theory, it is coming to be recognized that from practice comes the most urgent demand for its rapid development.

If necessity is the mother of invention, scientifically developed production is the mother of scientific research.”
Arthur Edwin Kennelly  (17 Dec 1861 – 18 Jun 1939) In ‘Scientific Research in the Engineering Schools‘, Electrical World (1920), 75, 151.

h/t Today in Science

+ + +

“There is one difference between theory and practice. Practice won’t let you forget anything or leave anything out. In theory, problems are easily solved because you can leave something out. I have found that an amazing thing. Any problem could be solved–if only it were different. Why the little problems should be so hard and some big problems should be so easy I never could understand.”
Charles F. Kettering in Speech “Lap-Welding Students to Life” collected in “Prophet of Progress: Prophet of Progress: Selections from the Speeches of Charles F. Kettering” Edited by T. A. Boyd

+ + +

“Entrepreneurship is neither a science nor an Art. It is a practice.”
Peter Drucker

+ + +

“We’re surrounded by the passion projects of the optimists of the past.”
Kevin Kelly (@kevin2kelly)

+ + +

“AI increases the demand for coordination work but doesn’t reduce the costs of coordination work. Watch for coordination challenges to increase in the near future.”
Lorin Hochstein (@NoRootCause) May-22-2026 tweet

+ + +

“I had an interesting experience at the beginning of World War II. We were running tests on a new type of propeller which was later used on boats during the war. We were running these tests in a little 25-foot boat up and down the Gulf Stream outside the harbor at Miami Beach. We always ran the tests in pretty bad weather, because we couldn’t find out anything in good weather.

“Two friends of mine, down there for a vacation, said to me one day: “Why don’t you go out and run that little boat up and down? Why do you have to go out and run that little boat up and down all the time in such bad weather? We think you’re crazy to do that. Why don’t you get somebody to do it for you?”

One of these friends happened to be an excellent golfer, and the other a good violinist. I said, “I never thought about that, but I’ll tell you what I’ll do. You fellows get somebody to practice your golf game and your fiddle for you. If you find that works, let me know, and I will get somebody to run that boat for me.”

You have to be part of the work. You have to live in the environment of the problem and then the facts and fallacies become apparent.”

Charles F. Kettering in a speech “Lap-Welding Students to Life” collected in “Prophet of Progress: Prophet of Progress: Selections from the Speeches of Charles F. Kettering” Edited by T. A. Boyd

+ + +

The threat to early founders is living in the future for too long.

Vision matters. It helps with hiring, selling, raising, and getting people to believe before the proof exists.

But vision is the engine, not the operating model.

“We’re going to win the market” is like saying “I’m going to win this chess game.” It may be true, but it’s useless for the next move.

The best founders I know keep the future in the distance and operate from the next measurable, pragmatic step, e.g. solve this specific problem for this specific niche at this scale. then let reality decide the next move.
Tom Ogrodzki in May-3-2026 Tweet

Ogrodzki is CEO of REDD, a provider of commercial real estate data for Central Europe. I included this quote because it captures the tension between a theory of the business and the practical need for mapping out next steps.

+ + +

“The human intellect is said to be so constituted that general ideas arise by abstraction from particular observations, and therefore come after them in point of time. If this is what actually occurs, as happens in the case of a man who has to depend solely upon his own experience for what he learns,—who has no teacher and no book,—such a man knows quite well which of his particular observations belong to and are represented by each of his general ideas. He has a perfect acquaintance with both sides of his experience, and accordingly he treats everything that comes in his way from a right standpoint. This might be called the natural method of education.

Contrarily, the artificial method is to hear what other people say, to learn and to read, and so to get your head crammed full of general ideas before you have any sort of extended acquaintance with the world as it is, and as you may see it for yourself. You will be told that the particular observations which go to make these general ideas will come to you later on in the course of experience; but until that time arrives you apply your general ideas wrongly, you judge men and things from a wrong standpoint, you see them in a wrong light, and treat them in a wrong way. So it is that education perverts the mind.

This explains why it so frequently happens that, after a long course of learning and reading, we enter upon the world in our youth, partly with an artless ignorance of things, partly with wrong notions about them; so that our demeanor savors at one moment of a nervous anxiety, at another of a mistaken confidence. The reason of this is simply that our head is full of general ideas which we are now trying to turn to some use, but which we hardly ever apply rightly. This is the result of acting in direct opposition to the natural development of the mind by obtaining general ideas first, and particular observations last: it is putting the cart before the horse.”

Arthur Schopenhauer in “Studies in Pessimism” Chapter: “On Education“

I think is true for entrepreneurial education that focuses on theory and does not provide practical experience.

+ + +

Some closing thoughts on what’s needed for an entrepreneur to succeed.

I think you need one or two insights that are contrary to conventional wisdom to set your theory of business. They should enable above average performance in an arena of your choice, or more accurately for a set of target customers. The day to day is focused execution of processes you have debugged and continue to improve. I think you have cultivate luck, which has more to do with building on your expertise and experience, honesty, helping others, and recognizing opportunities and some hustle. The more you explore, experiment, and learn from your experiences the more good luck you gather.  It reminds me of a quote by Branch Rickey I originally curated in August 2012, and used again in “Five quotes From Branch Rickey for Entrepreneurs” and “Cultivating Luck in Business Endeavors and Relationships.”

“Things worthwhile generally don’t just happen. Luck is a fact, but should not be a factor. Good luck is what is left over after intelligence and effort have combined at their best. Negligence or indifference are usually reviewed from an unlucky seat. The law of cause and effect and causality both work the same with inexorable exactitudes. Luck is the residue of design.”

Branch Rickey (quoted in Branch Rickey’s Little Blue Book: Wit and Strategy from Baseball’s Last Wise Man by John Monteleone)

Quotes for Entrepreneurs Curated in April 2026

I curate these quotes for entrepreneurs from a variety of sources and tweet them on @skmurphy about once a day where you can get them hot off the mojo wire. At the end of each month I curate them in a blog post that adds commentary and may contain a longer passage from the same source for context.

My theme for this month’s “Quotes for Entrepreneurs” is dreams, visions, and aspirations.

+ + +

“It takes a lot of courage to show your dreams to someone else.”
Erma Bombeck

I originally curated this in September 2010, noting that this is an early challenge entrepreneurs must overcome to move forward to be able to pursue their dreams.

+ + +

“The greatest achievement was at first and for a time a dream. The oak sleeps in the acorn, the bird waits in the egg, and in the highest vision of the soul a waking angel stirs. Dreams are the seedlings of realities.”
James Allen in “As a Man Thinketh” Chapter 6 (1902)

+ + +

Nick Bilton: Why did you start investing in space-related start-ups?
Steve Jurvetson: My interest in space started early, but for many years I could not find any space-related investments that really penciled-out for venture. That changed in 2009 when Elon Musk came to us with a big vision to explore Mars, while producing rockets at a fraction of a price and making space accessible.
From “A Conversation With Steve Jurvetson, Space Investor and Rocket Maker” (Mar-17-2014)

Related:

• Steve Jurvetson (@FutureJurvetson) Future Ventures / Flickr Blog
• Jurvetson accepted Churchill Club 2012 game changer award for Space-X (starts 10 min in)
• Steve Jurvetson Bats 1,000 at the Churchill Club 2008-17

+ + +

“In many schools today, the phrase “computer-aided instruction” means making the computer teach the child. One might say the computer is being used to program the child. In my vision, the child programs the computer and, in doing so, both acquires a sense of master over a piece of the most modern and powerful technology and establishes an intimate contact with some of the deepest ideas from science, from mathematics, and from the art of intellectual model building.”
Seymour Papert in “Mindstorms: Children, Computers, and Powerful Ideas” (1980)

+ + +

“The difference between a vision and a hallucination is how many people you can get to believe they see it, too.”
Gene Spafford

h/t Quotable Spaf

+ + +

+ + +

“The effort of the economist is to see, to picture the interplay of economic elements. The more clearly cut these elements appear in his vision, the better; the more elements he can grasp and hold in his mind at once, the better. The economic world is a misty region. The first explorers used unaided vision. Mathematics is the lantern by which what before was dimly visible now looms up in firm, bold outlines. The old phantasmagoria disappear. We see better. We also see further.”
Irving Fisher in “Mathematical Investigations in the Theory of Value and Prices” (1892)

+ + +

“As an armchair observer without a shred of economic acumen, I see a continuum of outcomes.”

Kyle Kingsbury in “The Future Of Everything is Lies, I Guess: Work“

A candid perspective in a thought-provoking essay on AI and Automation.

+ + +

“As entrepreneurs we pursue our dreams of improving the world. Sometimes I think our dreams can pursue us, that a vision of a better world can have the same effect as what the Scots call a geas. An obligation to use our talents toward certain goals that unlocks our power if obeyed and punishes if we don’t respect it.”
Sean Murphy in While We Pursue Our Dreams, They Can Pursue Us

+ + +

“Prudence is foresight and far-sightedness. It’s the ability to make immediate decisions on the basis of their longer-range effects.”

John Ortberg

+ + +

“The optimal state of inner experience is one in which there is order in consciousness. This happens when psychic energy—or attention—is invested in realistic goals, and when skills match the opportunities for action. The pursuit of a goal brings order in awareness because a person must concentrate attention on the task at hand and momentarily forget everything else. These periods of struggling to overcome challenges are what people find to be the most enjoyable times of their lives.”
Mihaly Csikszentmihalyi in “Flow”

Csikszentmihalyi outlines eight major components of a flow experience:

1. We have a chance of completing the task.
2. We can concentrate  on the task.
3. There are clear goals.
4. We get immediate feedback.
5. We are deeply involved in the task and forget our worries.
6. We feel in control of our actions.
7. Concern for the self disappears, yet paradoxically the sense of self emerges stronger after the flow experience is over.
8. The sense of the duration of time is altered; hours pass by in minutes, and minutes can stretch out to seem like hours.

+ + +

“Many situations in life are similar to going on a hike: the view changes once you start walking.

You don’t need all the answers right now. New paths will reveal themselves if you have the courage to get started.”
James Clear

Sometimes the decision to make a start is all the vision you need. This reminds me of a passage I quoted in “Burn Your Boats, Not Your Bridges” from W. H. Murray‘s 1951 book, The Scottish Himalayan Expedition.

“Until one is committed, there is hesitancy, the chance to draw back. Concerning all acts of initiative (and creation), there is one elementary truth, the ignorance of which kills countless ideas and splendid plans: that the moment one definitely commits oneself, then Providence moves too. All sorts of things occur to help one that would never otherwise have occurred. A whole stream of events issues from the decision, raising in one’s favor all manner of unforeseen incidents and meetings and material assistance, which no man could have dreamed would have come his way. I have learned a deep respect for one of Goethe’s couplets:

‘Whatever you can do, or dream you can do, begin it.
Boldness has genius, power, and magic in it.’”

W. H. Murray in his 1951 book, The Scottish Himalayan Expedition.

+ + +

“The hardest single part of building a software system is deciding precisely what to build. No other part of the conceptual work is so difficult as establishing the detailed technical requirements, including all the interfaces to people, to machines, and to other software systems. No other part of the work so cripples the resulting system if done wrong. No other part is more difficult to rectify later.”
Fred Brooks “No Silver Bullet: Essence and Accidents of Software Engineering” (1986)

Deciding what to build and specifying it requires vision. I blogged about “No Silver Bullet” and it’s sequel in Fred Brooks’ “No Silver Bullet” Revisited; I curated a quote from “No Silver Bullet in Quotes for Entrepreneurs Collected in February 2018, and curated the first two sentences of this quote in March 2017

+ + +

“The best way to keep something bad from happening is to see it ahead of time, and you can’t see it if you refuse to face the possibility.”
William Burroughs in “Rock Magic: Jimmy Page, Led Zeppelin, And a search for the elusive Stairway to Heaven “

+ + +

“Forecasting attempts to find the most probable course of events, or perhaps a range of probabilities.  But the entrepreneurial problem is to find the unique event that will change the possibilities, for the entrepreneurial universe is not a physical but a value universe.  Indeed the central entrepreneurial contribution and the one which alone is rewarded with a profit is to bring about the unique event, the innovation that changes the probabilities.”
Peter Drucker

This comes from a longer essay on this topic in the book Technology, Management and Society, p 110-112.  This was originally published in 1959 in the journal Management Science (available on JSTOR): Drucker, Peter F. (1959): Long-Range Planning: Challenge to Management Science. Management Science 5(3): 238-249.

h/t Russell Cameron Thomas

+ + +

“Strategy is a series of questions that lead to answers. It’s not a map—it’s a compass.”
Seth Godin

+ + +

“Evolution is a blind giant who rolls a snowball down a hill. The ball is made of flakes–circumstances. They contribute to the mass without knowing it. They adhere without intention, and without foreseeing what is to result. When they see the result they marvel at the monster ball and wonder how the contriving of it came to be originally thought out and planned. Whereas there was no such planning, there was only a law: the ball once started, all the circumstances that happened to lie in its path would help to build it, in spite of themselves.”

Mark Twain in “The Secret History of Eddypus”

An alternate perspective on vision, although in some sense he is describing the dominant eigenvector for a system: if you can design the system it will iterate to achieve the results you want.

+ + +

“There is nothing wrong in saying, “I don’t know.”
Don Valentine (quoted in “DTV” by Michael Moritz)

Admitting your ignorance can be the beginning of wisdom. It also encourages you to engage in foresight. “DTV” is a short read at 53 pages that offers a brief intimate perspective on Sequoia Capital’s founder.

+ + +

“Entrepreneurship isn’t about luck, it’s about vision, time management, creativity, determination and goals.”
Onyi Anyado

This is a good list, I like the inclusion of time management. I do believe that luck can be cultivated:

• Cultivating Luck in Business Endeavors and Relationships
• Increase Your Luck Surface Area To Get More Customers
• The Lucky And The Wise
• Entrepreneurs, Luck, and Silicon Valley
• Naval Ravikant On Avoiding Bad Luck
• Feeling Lucky Is Not A Strategy

+ + +

“A good question is never answered. It is not a bolt to be tightened into place but a seed to be planted and to bear more seed toward the hope of greening the landscape of idea.”
John Ciardi

I curated this originally in October 2012. In many ways a question may be more useful than a vision for an entrepreneur.

+ + +

“Some key inventions are business models rather than physical technology. Wheeled vehicles had existed for millennia before the brilliance of Pascal and the fluke of Baudry’s shuttle generated the idea of bus timetabling. Technological progress is of course immensely important. But our existing endowment of physical technology already holds untold thousands of applications that could transform our lives, just as Baudry’s buses transformed the lives of our ancestors.”
Samuel Hughes (@SCP_Hughes) in “The Invention of Buses“

This reminds me of an observation in “AntiFragile” by Nicholas Nassim Taleb:

“Implementation does not necessarily proceed from invention. It too requires luck and circumstances, and the wisdom to realize what you have on your hands. […] The Half-Invented. For there is a category of things that we can call half-invented, and taking the half-invented into the invented is often the real breakthrough.”
Nicholas Nassim Taleb in “AntiFragile”

I like this insight and Taleb’s quote that encapsulates it so nicely and used in “Mike Maples on Pattern Breakers,” and quote collections in July 2020 and November 2020.

+ + +

“Solving a problem simply means representing it so as to make the solution transparent.”
Herb Simon

A good vision statement can offer a similar level of clarity.

+ + +

“Our need to illuminate everything should be done thoughtfully since some things can only be seen in the dark.”
Beston Jack Abrams

+ + +

“Leaders talk about communicating a vision as an instrument of change, but I prefer the notion of communicating an aspiration. It’s not just a picture of what could be; it is an appeal to our better selves, a call to become something more.”
Rosabeth Moss Kanter in Leader to Leader (Summer, 1999)

+ + +

“Leadership is the capacity to translate vision into reality.”
Warren Bennis

+ + +

“Progress should mean that we are always changing the world to fit the vision, instead we are always changing the vision.”

G. K. Chesterton

In “Logic of Chance” (1866) John Venn writes about a man who drew a target circle around a bullet hole in a barn door to impress visitors.

“A man once pointed to a small target chalked upon a door, the target having a bullet hole through the center of it, and surprised some spectators by declaring that he had fired that shot from an old fowling-piece at a distance of a hundred yards, His statement was true enough, but he suppressed a rather important fact. The shot had really been aimed in a general way at the barn-door, and had hit it; the target was afterwards chalked round the spot where the bullet struck.”
John Venn “Logic of Chance” (1866)

h/t Bayesian Spectacles

+ + +

“Strategic planning does not deal with future decisions.  It deals with the futurity of present decisions.  Decisions only exist in the present.  The questions that faces the strategic decision-maker is not what his organization should do tomorrow.  It is, ‘What do we have to do today to be ready for an uncertain tomorrow?‘  The question is not what will happen in the future.  It is, ‘What futurity do we have to build into our present thinking and doing, what time spans do we have to consider, and how do we use this information to make a rational decision now?'”

Peter Drucker in Management: Tasks, Practices, Responsibilities (1974)

+ + +

Image source: 123rf.com/profile_peshkov and 123rf.com/profile_alstanova

Why Spreadsheets Still Rule Finance and Why That’s a Problem

By Martha Ryan

In an informal poll of 20 former colleagues, now working at 20 different Silicon Valley firms, every finance department used spreadsheets to produce their monthly and quarterly reports. All their companies are at least $500M in sales and they have a host of expensive ERP and FP&A applications that keep track of their operations and finance activities. So how do they close the books? Analyze pricing? Explore merger opportunities? Plan for expansion? Collect budget inputs? Generate forecasts? Integrate newly acquired entities? Spreadsheets.

Survey results cited by The Saas CFO on LinkedIn cites that spreadsheets have a 73% market share for financial planning and analysis. Finance Director Arnaud Lemaire summed up the situation nicely in a comment: “[Spreadsheets] are winning because the data feeding the forecast is messy, the process changes every quarter, the person building the model needs to tweak things nobody anticipated when the platform was configured. FP&A tools work great when inputs are clean and stable. Most companies aren’t there yet.”

Spreadsheets are used and re-used for critical decision making and reporting. They are almost all extremely complex with 10s of thousands of formulas. In this post, I will detail the problems with spreadsheets, how they come to be so complex, why companies depend on them, and what might be done to address the multiple challenges they pose.

Why Spreadsheets Become So Complex

Complex spreadsheets are built by talented finance professionals who take pride in their ability to tap into their deep knowledge of obscure formula types and layer on a bit of code to produce amazing tables and charts that point their leaders to a decision. And then, in previous job markets, this talent moves on to another position, another company. In today’s job market, talent is laid off with executives expecting that AI will fill the gap. The smaller remaining team has two options: decipher the thousands of 10-line formulas throughout 30 worksheets, tapping into Claude or Chat GPT, and still spend dozens of hours figuring out how the workbook actually works. Or these teams use the analyses as is without really understanding them. They make changes and pray to not see the dreaded #error. Or worse, they get no #errors and assume they are correct, only to discover months later that a key column has been excluding key data because of a formula error.

The Fragility Problem

CFOs know that their teams depend too much on spreadsheets. They queue up for IT resources to add code to the ERP systems to meet their needs. It’s a never-ending queue because business is dynamic and the needs constantly evolve. Other crucial projects take up IT resources and CFOs continue to rely on spreadsheets. CFOs expect that AI will help. They are not exactly sure how, and still have a deep-seated confidence that AI is the solution that will ensure the integrity of the quarterly results, the new direction the forecast is pointing, or the proposed acquisition really is a slam dunk. In a survey two years ago cited in The Wall Street Journal, 40% of public company CFOs were not confident in reported results. AI has been evolving at a screaming rate since then. Will it help?

Have you thrown a complex formula into Chat GPT to understand what it is trying to do? It kicks out a line-by-line definition of how the parentheses are impacting the “if, then” aspects of the formula. Claude takes it a step further and reflects the column/row header names, so you don’t have to hunt these down yourself. If you are good at keeping track of 10 different moving pieces and how they interdependent, you can then move on to the next formula in a complex workbook of 50+ formula types. What neither solution will do is decode the institutional memory buried in the workbook.

Let’s say that AI plus “elbow-grease” helps you decode the spreadsheet function, its dependencies, and its weaknesses. The institutional memory aspects often address the “why is this so complicated?” and is assumed common knowledge by the creator. You back into this awareness when you do the decoding. Often, it takes a whole other level of questioning to understand the history, data weirdness, and rationale for why “it’s always been done that way.”

No matter the case, spreadsheet are still the norm for critical analyses. A team dependent on spreadsheets is dependent on tools out of its control, understood in detail by one or two experts. The team that relies on these analyses and experts for repeated, critical path analysis is teetering on a fragile base.

How Complexity Creeps In: A Familiar Story

Complex workbooks often start with a modest request. “Please provide an analysis of the current MSRPs and determine the characteristics of lost customers as a result of the price increase last month.” Assigned the project, an enterprising analyst collects the relevant raw data from company ERP sources, uploads them into a new spreadsheet, and then starts sum-if-ing, xlookup-ing, match-ing, and tallying the results in a summary sheet. They look funny. He notices there are some outliers skewing the results. He figures out how to identify and exclude the outliers. To keep the workbook ‘robust’ so he can re-import the same dataset from a different period, he excludes the outliers with a formula. To not overwhelm the workbook, he nests that formula inside an xlookup.

The summary reflects new totals and still something is off. Research reveals there are some duplicates – not sure how those got in there. Make a note to follow up with the IT team to fix the data. He nests another exclusionary formula. The summary starts to look reasonable, so he shares it with a manager. She points out that it would be great to see if the channel had any impact on the pricing. His data doesn’t include channels. He learns that channel data comes directly from the distributor via a PDF file. After scanning it, he realizes the distributor PDF uses the marketing part number instead of the manufacturing part number. The marketing number has a one-to-many relationship with the manufacturing part number so matching becomes tough. The analyst figures out a nifty way to pivot table the results. The pivot adds three more tabs to the workbook including a crucial step of copying values into the third sheet – he makes a mental note to flag the hard copy sheet.

The evolution continues with the manager’s manager adding a few suggestions. Within a few days, the analyst has created a 30-tab workbook with table functions, pivot tables, over 10,000 formulas, hard coded entries, all parsing raw data from multiple sources to provide the analysis. The next day, the analyst starts to cleanup, labeling the tabs, and column headers. He is interrupted to work on another critical deliverable and doesn’t pick up the analysis again for a few days. He spends 30 minutes refreshing his memory of what he did on Monday and proceeds to make the workbook more robust with those cool somewhat obscure formulas he learned online that are just right for this analysis.

The meeting with the CMO is next week and the analyst has time to put together some clever charts which makes the workbook even larger and requires a bit more manipulation – a second and third summary tab so the data will reflect in the chart in a way that can be best appreciated by the CMO. Unfortunately, the CMO was called out of town on an emergency and the meeting has been pushed out two weeks. The analyst moves on to other things and as the rescheduled meeting approaches, tries to remember where exactly he put that analysis. After a fifteen-minute search, he finds the most recent version in a weird folder on his laptop where he saved it because the WiFi was broken at his apartment and he wanted to do some additional work at home. He refreshes the data (since a whole month has now passed) and fortunately everything looks good.

At the meeting, the CMO is very impressed with the analysis. She asks if it can be the basis of a forecast template? Of course! The analyst returns to his cube and makes a bunch more changes. Pricing updates every six months, and the new proposed pricing won’t be available for another two months. He saves the analysis on the network, labeled with a naming convention that is sure to be understood.

A month later the analyst gets an incredible offer from another firm and leaves the company. Another month after the analyst leaves leave, his now-previous manager has an outstanding commitment to the CMO for a new forecast deliverable and no idea which version on the network is the right one. All the data needs to be refreshed, and the next analyst doesn’t realize that the downstream summaries depend on hard coding the results of that pivot table on tab 25 (who would do that?) The next meeting with the CMO doesn’t go very well.

The scenario where key decisions depend on spreadsheets affects more than the finance department. The planning department and marketing department and reliability team and product development folks all rely on manipulating data from multiple systems which don’t quite capture their current needs. The industries affected are similarly widespread – manufacturing, SAAS, services, consultants, and government. It affects large, medium, and small companies. The smaller and younger the company, the more likely they depend on spreadsheets. A small concern I know with $20M in sales, had a plan to transition its planning from spreadsheets to a demand planning tool until “the downturn”. They are still using those same complex spreadsheets to plan their now $10M company with only one person who understands the plan.

Why Companies Stay Dependent

To summarize: spreadsheets are relied upon because large systems are incapable of adapting to the rapidly changing environment. They are complex because the data are imperfect and analysts must create huge workbooks and nest data cleansing and merging with the analysis. The spreadsheet skills of analysts are self-taught and different analysts use different formulas and approaches that can be a mystery to fellow analysts. The logic and sequence are difficult to visualize because they are buried in cells and sheets without an overall map. Once created, spreadsheets are saved in a range of shared and individual folders. Finding them requires diligence in naming the file, naming the folder, and managing versions. Limited time and resources mean that critical documentation details are frequently an afterthought.

Spreadsheets certainly have their place. What else gives you the flexibility to build on the fly, manipulate data, and summarize in tables and charts? However, as soon as you start using them repeatedly for critical decision making or reporting, you encounter myriad pitfalls of their fragility. When Arnaud Lemaire says that “most companies are not there yet” with clean data and stable inputs, I wonder what company is so staid and methodical? What analyst works in a predictable environment with such long sightlines that IT can be called upon to re-configure the ERP and finance systems to meet unanticipated needs nine to twelve months in the future? Dynamic describes all markets and the companies that address them. Even the government is rapidly evolving. Layer on budget and resource constraints in IT and among the data analysis teams, and spreadsheets soon become the analysts’ go-to for close, forecasting, pricing analysis, budgeting, integration, etc. These uses are constantly evolving and underpin critical decision making. Hence, the inherent fragility becomes a significant corporate risk.

This begs the question, really? After all these years, that’s what we have? Yes, that is the current situation. How does a CFO or finance manager deal with this reality of fragile dependency?

What Can Be Done: Practical Steps Forward

Getting your company’s data in order is a critical, if unglamorous, first step. AI can help find data anomalies and be enabled on a continuous search provided it is well trained on the typical examples of errors. A key source of institutional memory can be replaced with systematic corrections and overhauls. By eliminating the formulas in a spreadsheet that address data corrections, their complexity could be reduced by maybe 20%.

Reengineer with Structure (SIPOC)

Reengineering the spreadsheets periodically is another important step. With your AI assisted clean data, arrange your worksheets leveraging the concept of SIPOC:

• Supplier,
• Input,
• Process,
• Output,
• Customer.

Identify the data sources and what fields you need from each. That takes care of the Supplier and Input.

The Process can be a bear. Often first built with random improvements, on short timelines, workbook formulas get longer and longer.

1. Take a step back and understand the overall objective.
2. Figure out the simplest way most of the analysis can be accomplished.
3. Keep a list of the corner cases and don’t embed them in the general analysis. Rather, segregate them in well-labeled sheets with simple and, if necessary, numerous distinct formulas instead of long complex ones.
4. Consider simplifying by using models to address the occasional situations. For example, buffer the results with fixed percent adjustment, or group small outliers together as a single entity.
5. These techniques work well in future looking analyses where there are many unknowns anyway. For both general and corner case treatments, be sure to document the processes.

Outputs are your results, they go hand in hand with your Customer.

• What is your team looking for in the results?
• What questions are they trying to answer?
• What insights do they seek?

In defining your results, KISS. As the workbook creator, you are deeply invested in the analysis. You know all the corner cases and all the pitfalls. Chances are this knowledge will NOT clarify the results. Frame the output as primary objectives (no more than 3) and secondary objectives (your long list of details and exceptions). Restrict the graphs and tables to the primary objectives. If someone wants to deep dive, design the results of the secondary objectives for those who know spreadsheets well and can navigate the data leveraging filters and pivot tables.

Spreadsheets – The Strategic Risk and Opportunity

None of the ideas I have outlined here are new or rocket science. They reflect decades of experience and observation learned from the value of discipline. The fundamental concept is to think of workbooks as a process. By cutting out wasted steps and complexity from the spreadsheet process, the streamlined workbook will help leverage AI. It will also help transition to ERPs and analytical software. The most essential motivator, however, is the risk a team incurs by NOT being disciplined, by NOT simplifying, and by NOT harnessing the institutional memory. In the short term, the results can be wrong and/or lead to bad decisions. In the long term, it will guarantee missing the benefits of the AI train.

About Martha Ryan

Martha Ryan is a co-founder of zValuate, a startup that provides analysts a tool to bend data and answer complex queries. Visual, fast, accurate, and collaborative are key objectives. Previously, she spent 25 years in finance and accounting at two Fortune 500 technology companies, deep in spreadsheets. Always striving to make things better, she earned a Lean Six Sigma Black Belt and led 100+ cross-functional
process improvement initiatives. She and a few colleagues invented zValuate to address the challenges of spreadsheets that frustrated them as analysts.

Related Blog Posts

• Scale by Understanding Your Value Stream: Slides and Video Recap Martha Ryan and Terry Frazier offered a briefing on how to “Scale by Understanding Your Value Stream” at the Lean Culture on Feb-25-2021.
• Consider What’s Changed And What You Bring To An Opportunity
• Experiments Vs. Commitments
• How to Drive Innovation and Meet Commitments
• Slides from “Limits of I’ll Know It When I See It” Talk at SFBay ACM
• The AI Gold Rush: It Still Takes Teams
• Intrapreneur Mindset and Key Skills

Mark Stiving walks through the Six Laws of Value as a practical map of how buyers actually make decisions. Not how we wish they did. Not how sales decks assume they do.

Each law explains a specific point where buyers hesitate, defer, or quietly disengage. When those breakdowns happen, price becomes the scapegoat.

Mark explains that pricing success depends on understanding buyer confidence, not just features or cost. Buyers make decisions by predicting future outcomes and choosing the option they feel most confident about. Confidence is driven by three factors: payoff (value), probability (likelihood of success), and anticipated regret (risk of failure).

Value comes from solving real customer problems, yet buyers often express solutions instead of underlying issues. Sellers who clearly articulate customer problems—better than the customer can—build trust, increase confidence, and raise willingness to pay. Most deals fail due to lack of commitment, not competition.

Mark Stiving on Pricing Problems

About Mark Stiving

Mark Stiving is a pricing expert, speaker, and author who helps companies make more money by figuring out how buyers actually decide what to pay, then aligning products, packaging, and pricing with that reality. He pushes hard on value-based and context-driven pricing and spends very little time on academic theory and a lot of time on what moves revenue and profit in the real world.

He’s best known for simplifying complex pricing concepts into frameworks that executives and teams can actually use. His work shows up in books, bootcamps, boardrooms, and consulting rooms, often with companies that have strong products but weak clarity on how to justify and capture their value.

The unsugarcoated summary: Mark helps organizations stop guessing at prices and start getting paid for the value they already deliver.

Few voices resonate with the clarity, insight, and expertise of Mark Stiving. His profound understanding of buyer behavior and his unique ability to distill pricing strategies into actionable insights have cemented his reputation as a sought-after advisor and instructor. His firm, Impact Pricing, works with private equity firms and their portfolio companies to focus on customer value, optimize pricing, and maximize valuation.

He hosts the popular Impact Pricing Podcast, has a Ph.D. in pricing from U.C. Berkeley, and is the author of four books:

• Impact Pricing – Your Blueprint for Driving Profits (2011)
• Win Keep Grow – How to Price and Package to Accelerate Your Subscription Business (2021)
• Selling Value – How to Win More Deals at Higher Prices (2022)
• Instant Profits – How to Raise Prices Without Losing Customers (2024)

The AI Gold Rush: It Still Takes Teams

Teaser quotes

Sean: We’re in the age of gentleman science. They’ve deployed these models, but they don’t really have a use case. It’s kind of like they’ve deployed new instruments and now we’re learning how to play them.
Jeff: It’s not going to happen overnight. And how is that change going to be proliferated into your organization? How are you going to leverage this new technology? It’s teams that develop products, not just individuals.
Sean: Westinghouse would pick a hard problem to solve that was the first example of a new kind of use case. The cost reduction drives more use and then more use leads to novel structures.
Jeff: Okay, time out here. This is how this thing works. This is how we want, this is how we need to implement it going forward.
Sean: NVIDIA is not talking about laying people off. Right? NVIDIA’s probably embraced this and they’re doing some amazing stuff and they’re not saying yes. And what we hope next year is to lay off 20% of our staff.

Edited Transcript

Norbert: This a conversation with Sean Murphy and Jeff Allison, both of whom have spent decades in Silicon Valley high tech. We break down what AI is, what it’s not. And how organizations can adopt it without losing the craft of engineering. I’m Norbert Korny, let’s jump right in.

Sean: I’m really excited to be here, to be in a conversation with some other guys that actually know what they’re talking about, so that I can, by association, at least appear as if I belong in the room. I’ve worked on introducing various forms of automation, computerization, and digitalization to teams and organizations since 1980.

I think that AI provides new affordances and new capabilities. But there are a number of useful rules of thumb for innovation that will still hold. We are in the process of discovering what these new possibilities can do for us. But my focus has always been, going back to Doug Engelbart, the augmentation of human intellect, not the replacement. The trick is to augment people who are good at their jobs to help them develop and extend their expertise.

This idea of technology change, managing the introduction of new capabilities into organizations, is something that I’ve worked on for small firms and large firms for a while. In 2003, I started SKMurphy, and I now work with early-stage technology companies, helping them navigate the introduction of new products and methods to business customers.

Jeff: My name is Jeff Allison. I’m so glad that you’re pulling this together, Norbert. There are so many conversations happening around AI, and so many that need to happen.

Every day, I open my newsfeed and read that AI is going to do this, AI is going to do that, and wonderful things are going to happen. And I think they will, but the conversations need to get into more detail about the realities of AI and how it can help.

I have spent more than 30 years in the Valley. The Valley is just an innovative hotbed. It’s full of very creative people who have created fantastic technology. There are so many unsung heroes. You walk down the street and may meet someone who was involved in developing IBM’s ATM infrastructure years ago. They were able to do that because they were supported in the challenges they faced. Their organizations invested in tools and methodologies to enable what they developed individually and in teams.

I think AI is going to help push innovation to much higher levels, but it’s not because they’re using AI; it’s because of how they’re using it and how we create infrastructure to enable that. It’s not just about saving cost; that’s not the best way to look at it. It’s about how to help people do their best in a shorter period of time in a way that can be absorbed by the overall team.

Individuals come up with great ideas, but teams make products. Companies invest in teams. I have not mastered the new AI tools, but I do know how to introduce new technologies into organizations and what that entails. And it’s not as easy as, well, we’ll just bring this thing in, and people will start using it, and things will go great. There are all sorts of issues that need to be figured out. And I think being able to share some of my experiences with how difficult that has been over the years will help introduce this new technology into the development process.

“The future is already here–it’s just not evenly distributed.”

Norbert: I know you like the William Gibson quote, “The future is already here–it’s just not evenly distributed.” What do you think that quote gets right about AI in 2026?

Sean: It’s spot on. The new path forward only becomes clear in hindsight.

There’s an article going around that we’re in the age of gentleman science, where individual kind of amateurs are creating breakthroughs using chatbots and AI tools. And I think the major AI companies, they’ve deployed these models, but they don’t really have a use case. It’s kind of like they’ve deployed new instruments and now we’re learning how to play them. And so a lot of people are exploring the landscape, And I don’t think we’ve quite figured out the implications of what it means. And so I think what it gets right is that we’re back in an age of kind of individual exploration, individual tinkering. And we’ll look back and realize that something that happened six months ago or nine months ago, or maybe a year and a half ago, actually was the branch of the tree that we’re following. But right now we’re in this forest and it’s hard to tell.

It’s spot on. There’s an article going around that we’re in the age of gentleman science, where individual amateurs are creating breakthroughs using chatbots and AI tools. And I think the major AI companies have deployed these models, but they don’t really have a use case. It’s like they’ve developed new instruments and now we’re learning how to play them.

A lot of people are exploring this new landscape, but I don’t think we’ve quite figured out the implications or what it means. We’re in an age of individual exploration, individual tinkering. At some point, we’ll look back and realize that something that happened six months ago or nine months ago, or maybe a year and a half ago, actually was the branch of the tree that we’re now following. But right now we’re in this forest, and it’s hard to tell.

What’s Driving the AI Gold Rush?

Jeff: People who are doing new things or just starting out in their endeavor to create a new product or deliver a useful innovation, pick up new methods naturally. But 90% of us continue to stick with “the old way” because it’s proven, predictable, and we understand it. We need to talk about how those folks can get on board with this new product development paradigm. And I don’t think there’s enough being talked about that.

Also, the rate of change in AI tooling is still so high that new methods are evolving faster than they can be documented and debugged. So this transition from explorers and early adopters won’t happen overnight.  A lot of what’s being described is experiments, pilot projects, or proof-of-concept. There is much less about how a successful pilot is proliferating as a proven methodology in an organization. Much less about how to use AI technology to improve your methods and provide better products for customers. I think that’s where we’re at.

Norbert: Is the Gold Rush really about efficiency, innovation, or a competitive arms race?

Sean: I think our sense of a Gold Rush is driven by competition and a fear of obsolescence. I think there’s been a lot of messaging that “Everything you know is wrong” or “Everything is going to change.”

Lately, I’ve been fascinated by the introduction of electricity into both industry and daily life. In some ways, it was a struggle between Edison and Westinghouse. Edison was a tinkerer who would try many things in parallel. One reason I think Westinghouse ultimately set the direction toward an alternating current architecture was that he chose hard problems and solved them for industrial use. This was formed by his early sales to the railroad of an air braking system that had to work reliably.

When he wanted to test his AC dynamo paired with an AC motor, he went to a remote mine in Colorado and ran it all winter, debugging problems as they came up. His approach was to solve a real problem with a satisfactory level of performance. That’s missing from much of today’s embrace of AI. Many announce tiny incremental breakthroughs, typically based on contrived benchmarks, not real-world outcomes. Everyone wants to be five minutes ahead, driven by a massive fear of missing out.

How do we Define and Adopt a New Paradigm?

Jeff: New tools, whatever they may be, can help us come up with very creative ways of developing products. We have been doing this for decades in the Valley; it’s great. The more you can automate and enable individuals to be more productive, the more wonderful it is.

The challenge is that teams develop products, not just individuals. Whenever you’re bringing in something new, you have to consider how it will work within the team.
It’s great to become more efficient in one area, but that may account for only 10% of the overall product development cycle. It’s an improvement, but if it slows down 90% of it, then it’s not really a win. It takes a team to develop today’s complex products that deliver value for customers.

Part of what I worry about, and what I want more discussion on: how do we leverage this technology so that the team gets more efficient, the team gets more competitive and the team is able to innovate beyond current constraints and not just accept them.

When we would develop algorithms or integrated circuits to perform critical tasks, we would then review results and say, “Okay, this is good, but we need to make it better.” So, if this technology allows us to make things better, I think it’s a great thing. But how do we do that at a team level, not just at an individual level?

I think the true Gold Rush is about the teams that really embrace this technology and are able to use it at the team level to collaborate and to develop product, they’re going to win. The guys that just take it and say, okay, this makes me more efficient, and they’re in a little box doing it, They’re not going to win here. They’re just going to get more efficient at what they do and it’s not going to affect the overall product.

Norbert: Yeah, and it seems like the adoption is driven by the individuals, unlike previously for enterprises, even better yet, the army before then. And it acts like a force multiplier. To be best, you need to be good at the beginning.

Jeff: Yes, you do. And we observed this when we were putting automation into the hardware development process for developed systems on chips. We got really good at doing specifications, but then we realized that we were really good at creating millions of gates, but now we have to get really good at testing them. So then we had to bring more people on and develop methodologies around that.

So I think this is what’s going to happen with this AI. AI is going to allow us to move very quickly in certain areas. But this will uncover blind spots that we will need to address. We’ve moved really quickly here and were really innovative there. But what does that mean to the overall development process? What else do we need to do to get this thing to actually realize everything that is promised?

Sean: Couple of quick thoughts. In the beginning, most technologies get adopted because of a cost reduction of some sort, or they come in as toys or kind of an artisanal tool.

I think a lot of AI has been sold to the CFO, by promising to cut costs by allowing you to lay people off. The most recent example is Block just laid off 40% of their people. I think that’s a smokescreen for years of bad decisions on their part. When you actually look at who’s getting laid off, it’s a lot of people with expertise. So I don’t buy that story.

The next step after cost reduction is more use, and then or use drives the creation of novel structures. Right now we see time savings at the individual level and thus more use.

Norbert: I agree.The story Jack Dorsey is running with is that they built smaller teams to make them more effective. Is this is the best thing we can do right now, because an individual wouldn’t make that big a difference in an enterprise? Small teams, maybe yes, but if you got 10,000 people, is it enough?

Individual Insight Can Enable Breakthroughs

Sean: I think an individual can have a significant insight that leads to an architectural or a novel combination breakthrough. The breakthrough always comes first in one person’s head, or maybe two people working in an intense collaboration. But then, to actually ship a product or deliver an industrial-strength or highly reliable processor, normally requires a cross-functional team of four to twelve to look at it from all the right angles.

If we go back to the Gold Rush, you can trace an evolution of individuals panning for gold—which is an artisanal model—to industrial-scale mining and extraction methods. Today, I think we’re still in the panning-for-gold phase. I don’t think we’ve even figured out how to facilitate a team of six to twelve.

Jeff: Well, this technology was initially developed for consumer or individual use, not industrial. I don’t see much of enterprise or industrial focus on what teams of engineers or scientists need.

Sean: It’s hard to track everything that’s going, but the Perplexity guys seem to be passing a test of giving answers with substantiation with links. As opposed to providing an answer that you have to treat as if it was delivered by an oracle. So Perplexity is at least a half-step towards deploying something that could be part of a reliable business process.

Jeff: That’s a big step from a consumer-based product to company-base. We’ve all done that. If anyone’s trying to bring technology into their organization and it works well for an individual, there is still a long process to get it adopted by one more more teams.

It requires a higher level of reliability, support, and well-defined infrastructure to enable all of that: standards, databases, defined development processes, revision control, methodologies, etc. These support infrastructure elements all take time to figure out and develop, regardless of whether they are for an outside vendor or an internally developed tool. There are many steps from a proof of concept to getting something widely deployed, making many people more productive.

Different parts of the organization may have different architectures and different cultures. An established business unit can have a big centralized architecture, but newer smaller business units may be more decentralized with small teams, some scattered all over the world. And then Different organizations have different architectures, right? They work differently and may benefit from different support and infrastructure models.

I think there’s a lot to be discussed here. And I’m happy, Norbert, that you’re doing this and trying to shine some light on some of these issues so that we can have these kinds of discussions and kind of work through it. Because I think some of these things are just not getting talked about enough.

Norbert: Is the software something where we’ve seen the biggest impact? Because from my point of view, it seems to me the software seems to be reinventing itself. We introduced the concept of shared libraries to reuse logic. AI recently introduced skills. Now you need a skill repository, and there are  “malicious skills” you need to scan for security threats. And this is a whole software development lifecycle you need to revisit. And we’ve been through this already.

Adoption Takes Months to Years

Jeff: We could draw lessons from earlier disruptive technologies for software development. How long did they take to go from individual adoption mainstream use. It wasn’t days or weeks. It was months and years.

I don’t know how long AI will take to become a reliable mainstream development environment for even small teams. But I think to Sean’s point where you’ve got people that are making decisions on this technology, they really don’t understand the technology, but they are just looking at cost savings. I can pay for an agent that can develop a product that costs X vs. 5-10X for a human being.

That’s a metric, but to me, it’s very short-sighted. It’s no way to run the business.
I think we are at a point where the business guys are running the show, but I’m hoping that the engineering and development organizations step in here and to say, “Hey, okay, time out here. This is how this thing works. This is how we need to implement it going forward, and this is what’s going to make sense for us rather than just counting the dollars.”

Sean: The business guys remind me of the early rapid early evolution of HTML. Netscape on a whim decides to add the blink tag or to implement cookies in a certain way. I think we’re going to look back on MCP and on some of these other things constructs and conclude that they should have never left the lab.

MCP is an interesting first effort, but it’s not a reliable, scalable protocol, and it throws away 90% of what we know about security and protocol design. I’m okay with the individual experimentation and allowing individuals to scout and report on what they were able to do. But because of this consumer focus and this fixation on replacing professionals, knowledge workers, or white collar workers, they really haven’t engaged the people who define processes, the methodologists who look at this in a structured way. And it doesn’t seem like those people are on staff at the major AI companies; they seem to hire people with different expertise.

Another interesting thing for me is that Jensen Huang at NVIDIA is not talking about laying people off. NVIDIA has deeply embraced AI and is doing some amazing stuff, but they are not saying, “Next year we hope to lay off 20% of our staff.” They realize they are in a Red Queen Race, running faster and faster just to hold on to their position in the market. They understand that the bar for acceptable performance will be continually raised.

Jeff: I think NVIDIA is airing this technology out internally and they’re probably developing methods, processes, maybe developing tool chains around this to enable them to develop their product. I think those tool chains and those methods will become more available to people over time.

This reminds me of design automation in the early days of hardware. The first Design Automation Conference was held in 1964, but it wasn’t until 1982 that they formally recognized independent vendors with an associated trade show. EDA was incubated by larger organizations for almost 20 years. IBM and other big companies developed technology internally, so they understood it.

And then, those technologies became more commercially available, creating a whole market. So I think that has to happen here too. And I think NVIDIA is probably taking the lead on it. And there may be more companies out there that are doing stuff behind the curtain, right? That, you know, they’re not talking about right now, but they’re maybe saying, okay, we’re going to develop this next generation of product or this product, you know, we’re going to use an AI model on this product and we’re going to air it out. And once we figured out how it all works, then we’re going to make it more mainstream in our organization. Maybe that’s going on. That’d be great. It’d be great to talk to people that are doing that. The AI automation gurus of the future, right? Where are they and what are they doing?

Sean: It seems like there’s two tracks running here. So, I agree it’s very possible that there are organizations that are getting value and deploying this to make internal processes or design processes more efficient, and they’re getting big advantages that they’re not disclosing. Because most of the stories that are being told seem to be told from the perspective of, “Don’t worry, we’re not falling behind. Here’s some examples of how we’re using AI.” But when you dig into most of those stories, it puffery.

We are still in the early stages. The Design Automation Conference was a community of practice that allowed competitors to compare notes and learn. People would not give away everything they were working on, but there was still value in solving common problems.

I remember watching a YCombinator LightCone podcast where they essentially concluded, “We think we can solve electronic design kick the currently EDA industry in the ass in a year from now.” Not so much. The semiconductor industry and the EDA guys have been running a Red Queen Race ever since they passed Moore’s Law. They are incorporating this technology. We haven’t so far seen anyone come out who’s made current practices obsolete. I think the current vendors are obsoleting their current tools, but that’s different from some outsider using this to destabilize the industry. All the majors now have a VP of AI. Everyone on the semiconductor side, the systems houses, and the design tools are all looking at this. I think it will be like electricity: It’s going to raise the level for everyone.

Jeff: VP of AI? Good.

Sean: You asked where we’re seeing the biggest impact. I think right now it seems like it’s amongst experienced developers working on greenfield designs. That to me is kind of a splinter or a scouting activity. There’s some changeover point where you’ve got a working proof of concept and now you actually have to harden it and deploy it and support it. And that’s not going to be normally just one person. I don’t know that we’ve got the models for that yet.

Norbert: Yeah, but there’s the rush for AI to replace humans, especially the juniors. Do you see the impact on the job market in Silicon Valley?

Sean: That’s absolutely going on.

Related Blog Posts

• Mark Bennett on Using Claude Code for Application Development (Q&A)
• AI in Action: Practical Automation by Alex Panait (webinar)
• Matt Trifiro on Lessons Learned using AI for Marketing (webinar)
• Andrew Shindyapin: AI’s Impact on Software Development (Q&A)
• Alex Panait on Current Trends and Possible Futures for AI (Q&A)
• Time to Market S01 E03 – How Will AI like ChatGPT Impact B2B Founders? 

References

Gentleman Science

• Sean Goedecke “We are in the “gentleman scientist” era of AI research” (Oct-18-2025) “Many scientific discoveries used to be made by amateurs.”
• Garry Tan “AI is in its Gentleman Science Era”

YCombinator LightCone Podcast: Why The Next AI Breakthroughs Will Be In Reasoning Not Scaling

“One thing that really stood out about o1 in particular—if you read one of the papers talking about it, so capabilities and potential for the future—it talks about how it does really well in chip design.[…]  At some point the AI will get good enough to just like design chips better than like humans can, and then it will just like eliminate one of its bottlenecks for like getting greater intelligence. And so it feels like that’s already kind of like we’re on the pathway to that in a way that we just weren’t before.

SKMurphy Take: AI may have a strong impact on electronic system design but it’s unlikely to enable a startup to obsolete the established players. See for example

• “How to Plan Agentic AI Deployment for Chip Design.”
• AI Growing Impact on Chip Design

Image Credit: 3 views of Antelope Canyon (c) Kevin Murphy, used with Permission

This post was republished on LinkedIn at https://www.linkedin.com/pulse/ai-gold-rush-still-takes-teams-sean-murphy-lua4c/

Hope for the best, prepare for the worst

Some thoughts on the value of preparing for a range of outcomes from your current course of action.  I think the most famous version of this advice comes from a song by Mel Brooks.

Hope for the best, expect the worst
The world’s a stage, we’re unrehearsed

Mel Brooks lyrics to The Twelve Chairs, “Hope for the Best, Expect the Worst” (1970)

I think the best version comes from Zig Ziglar, urging a flexible response to a range of outcomes.

“Expect the best.
Prepare for the worst.
Capitalize on what comes.”
Zig Ziglar

The oldest versions I could find come from Benjamin Disraeli and Mary M. Bell, but this may have been a proverb in circulation in the 19th century.

“Hope for the best,
prepare for the worst,
and take what God sends.”
Mary M. Bell in “Seven to Seventeen; or, Veronica Gordon” (1873)

“I am prepared for the worst, but hope for the best.”
Benjamin Disraeli in “The Wondrous Tale of Alroy” (1845)

Tony Hsieh offers a more tactical prescription in his book “Delivering Happiness” that also applies to cash management for bootstrappers:

• Always be prepared for the worst possible scenario.
• Play only with what you can afford to lose.
• Make sure your bankroll is large enough for the game you’re playing and the risks you’re taking.
• Remember it’s a long term game. You will win or lose individual sessions, but it’s what happens in the long term that matters.

Tony Hsieh in his book “Delivering Happiness” in the chapter on “Poker”

I used this in Texas Hold’Em as a Model for Technology Startups.

“What you prepare for, what you hope for, and what you expect are often three different things.”
John D. Cook (@JohnDCook)

Make affordable loss bets

“Look at the Underside First: Legions of people are paid large sums to promote the positive aspects of commercially available products. Very few people earn their daily bread by pointing out malfunctions, bugs, screw-ups, design failures, side-effects and the whole sad galaxy of trade-offs and failings that are inherent in any technological artifact. To counteract this gross social imbalance, a wise designer and a wise critic will make it a matter of principle to look at the underside first.”  Bruce Sterling in Viridian Design Principles (1998)

There is much more at the Viridian Design Site. In “Constructive Pessimism,” I observed that you have to be willing to acknowledge the possibility of problems and look for them to be able to prevent or at least manage them. Many entrepreneurs who are naturally optimistic make a serious mistake in discouraging pessimistic thinking instead of putting it to good use. The clever utilization of constructive pessimism is one of the keys to success.

Postscript: Optimism vs. Confidence

Axiom 9: Optimism is expecting the best, confidence is knowing how you will handle the worst. Never make a move if you are merely optimistic. Know how you will handle the worst.”
Max Gunther in Zurich Axioms

Related Blog Posts

• Constructive Pessimism
• SKMurphy Perspective Counterbalances Excess Pessimism or Optimism
• Managing Recurring Problems In Your Startup
• Startup Uncertainty At The Very Beginning
• Making Business Decisions in Uncertain Times
• Q: How Do I Plan for Pivots or Even Shutting Down My Startup?