I like to create models and frameworks for work I’m doing. When I’m working on a conversational AI/ Intelligent Agent (IA) system I have some categories of design so that the creator of the product/ use case/ design can organize our thought. This also helps with collaboration with software engineers.

Communication
How will your agent communicate with humans and other agents?

Skills
Also generally defined as “workflows”. What tasks will the agent be able to accomplish and how will it learn how to do them?

Memory
Each agent needs to access data/ memory. Is this data store proprietary to each agent or is it shared across agents (or, some mix)?

Personality
In order to more effectively interact with humans, what personality will the agent have? Will the agent have one personality or will you customize the personality for each human it interacts with?

Privacy
The agent should be aware of personal data and also privacy hierarchies so that the right information is shared with the right people. It’s possible to have information that is accurate but non-shareable in different contexts.

Advanced Contextual Awareness
The agent should be aware of as much context as it can get to fulfill its mission.

Behavior
The agent should understand its communication channel context in order to behave properly in different contexts.

Rob and I go way back and this is the type of thing we talk about over beers. I feel obliged to add some notes of my own. I wish I vehemently disagreed on something. That would perhaps make for a more interesting post.

I will also give the caveat I give in talks at AI conferences — I don’t claim to be an expert on AI trying to innovate. I’m an expert on innovation trying to work on AI.

Here are 5 key, interesting things to discuss around AI:

1. AI innovation is limited

Innovation is one of the things I’ve spent time studying and trying to intentionally create across many products and technologies. The media likes to create clickable news across a wide array of topics (I suppose all topics).

So, when I talk to experts (no matter the expertise) they have some level of eye-rolling concerning the media. Nuance gets oversimplified.

There’s a lot more to come: “AI fields like symbolic logic, evolutionary algorithms, and others have hardly been touched, and even for neural nets much of the work has been researchy, and is difficult to translate into applications.” (Inside AI)

As with most technologies, I advise people to lower expectations on what can be accomplished in 2 years and raise expectations on what they think can be expected in 10 years. The reason this phrase is often true is that humans are imagining improvement as linear when it’s actually exponential. And, this is exactly what it looks like when you map a linear improvement to an exponential curve. This of course assumes we’ll continue an exponential pace.

I advise people to lower expectations on what can be accomplished in 2 years and raise expectations on what they think can be expected in 10 years

2. AI hardware is less talked about but important

I attended Mythic’s B round celebration. And, I attend events where we talk about quantum computing. This will be a fun one to watch from the sidelines and making investments. It will be a massive revolution. Not much more for me to contribute here. I’m learning and watching. I recommend others learn and watch as well. This is a great area.

3. “AI is X”

I was one of the people propagating the imperfect “Data is the new oil” metaphor years ago. I spoke about it at conferences and followed up with Medium post. My original pitch was that if you look at the very largest companies, we’ve moved from oil-based companies to data-driven companies as the leaders in creating new value. Data replaced oil as the driver of the value.

Also, oil comes from raw materials that need to be refined in order to be useful. Having done early big data, then AI projects at the corporate level, I noticed that’s where corporations get stuck. They are sitting on lots of data, but it needs to be refined. There are other areas (tangibility and fungibility for example) where the metaphor doesn’t work. So, I wouldn’t read too much into it beyond a simple story I tell to help executives understand what they are up against. It ties in well with our conversations about how all companies are becoming software companies going back 20 years.

“AI is the new electricity” falls into this category. I just interpret it as saying it will be ubiquitous and also enable new capabilities. Electricity did those things. As a business/ market study, it probably won’t work.

The interesting thing about the “AI is electricity” metaphor is that AI will be ubiquitous. In 10 years AI will probably morph into just being a category of “software development”. It will be as ubiquitous as microchips. Notice that no startups pitch themselves as “we run on microchips” today. It’s ubiquitous and obvious. The tech will be readily available.

“AI is the new electricity” falls into this category. I just interpret it as saying it will be ubiquitous and also enable new capabilities. Electricity did those things. As a business/ market study, it probably won’t work.

What this really comes down to is that there may not be any good metaphors for what’s happening. Technological civilizations only make all their software intelligent once, and it doesn’t look exactly like anything else before or after. That’s why AI is interesting. The arc of technology will happen once everything will be different. That’s what makes it interesting.

What this really comes down to is that there may not be any good metaphors for what’s happening. Technological civilizations only make all their software intelligent once, and it doesn’t look exactly like anything else before or after.

4. New workflows require a culture change

This is where innovation gets slowed down. The technology pioneers have a vision of where they want to go. And, culture changes much more slowly.

A really intelligent AI developer who was doing amazing things called me last week to chat about AI use cases. There are tons of developers ready to work on things. And in Silicon Valley, many use cases (the blatantly obvious use cases get more investment and therefore more technical attention). He wanted to know some ideas about how to think about opportunities. I told him to look beyond the technological trends and also the societal/cultural.

Bill Gross mentioned that timing is the number one issue causing the success or failure across IdeaLab companies. Timing is a mix of “can you build it” and cultural timing. When you get to the point where you know you can build it, the timing issue comes down to society and cultural timing. I think that internal corporate cultures are like overall social culture changes. They go in waves, and they impact timing.

I’ve created software products since I was 12. I’ve done it professionally for 20 years. In those 20 years, I’ve probably launched 30 products. The vast majority actually were correct on the idea. But, none had perfect timing (too early by 5 to 10 years). So, I actually worry less about having the right idea now and worry more about whether I’m inputting it into society and corporate cultures at the right time. If it takes 6 years longer (as often happens) it might not fit the venture model. (I have an alternative model idea, but that’s another discussion.)

The workflow changes I see are around workflows for knowledge management and workflows around augmenting the work product itself using pattern recognition. Both types improve the work product and effectiveness of the work so much that there’s no stopping them.

Now, we’re just down to timing. Once we get them working, it still comes down to timing based on when society and companies will accept them — and acceptance means people will need to change as well.

The big opportunities don’t just change tech around how humans live. The big opportunities require human behavior change as well. Which we need to be aware of, but it’s difficult to predict.

The big opportunities don’t just change tech around how humans live. The big opportunities require human behavior change as well.

5. Bias and optimization dangers will happen way sooner than AI sentience issues

The question that people (and the media) like to talk about is AIs getting so smart they take over. I love to chat about it. The bigger problem that will actually happen sooner is the optimization problem.

The “optimization problem” I would define (as Rob gives an example) as the problem caused when we create systems that optimize at the macro, not at the micro (just you) level that we’re used to. A lot of AI systems will get their sustainable value from macro decision making (you become an input and also benefit from the output of the system). The problem is created when you don’t understand how you got “optimized” vs the collective. When are you getting optimized for your personal benefit vs. being optimized for the collective benefit?

When are you getting optimized for your personal benefit vs. being optimized for the collective benefit?

There’s an inverse correlation between AI talk and AI development. I talk to a lot of really smart people in Silicon Valley daily. The closer to the code people are less concerned about the dangers of sentience and more about bias. I want to add optimization to that list as we build systems that optimize across the population. Bias and optimization problems are here now.

In Information Technology and tech startups, we talk about “systems of record”. It’s a system that is the source for particular data. That “system or record” is the ultimate source of truth. It’s the canonical record of data.

We often talk about these systems as the place a single company puts data they can refer to later as the single source of truth. A modern company will have a bunch of systems of record including products from companies like: SAP, Microsoft, Oracle, Intuit, Workday, Salesforce, Atlassian, etc.

As we move into the federated and distributed world of IoT and AI, new companies will need to be formed to store and process common data at the next layer of abstraction above these names. This is the universal system of record.

As we move into the federated and distributed world of IoT and AI, new companies will need to be formed to store and process common data at the next layer of abstraction above these names.

This is the universal system of record.

The key to this model is that the value of the company or system comes from creating and managing a common pool of data shared among independent institutions.

The value comes in two parts:

1. They are a place to store and share common data
2. They are a place where common wisdom can be derived from the stream of shared data

This model won’t make sense for some companies, use cases, and data. But, as the value of large amounts of data becomes greater, the inventive to share data to create new use cases will become too great.

And, creating your own features based on your own data (and not being a customer of the USR company) won’t be worth it. The networked data from your industry will be more valuable than your own siloed data.

The networked data from your industry will be more valuable than your own siloed data.

What does a universal system of record (USR) look like?

The characteristics of a USR are:

1. Multiple institutions/ company have useful data they are willing to share
2. There is value to the data across an industry or wider group
3. One (or small group) of companies form to store common data
4. Depending on the industry and use case it may also enable a learning loop by processing the data it has access to

Examples of current universal systems of record

Business communication data: MessagePath
The application I’m building with several other people will create a universal system of record for business communication patterns that will benefit all customers. On its face, the mission of MessagePath is to facilitate global business communication. But what’s different about us is that we’re going to be the canonical store of optimized business writing. Individual companies won’t need to keep up with the nuances of the language of business for marketing, sales, support, industry, and legal nuances. They will just use (or plug-in via API to) MessagePath.

We have 4 API customers in the pipeline who know that the only way to get the language patterns that MessagePath supplies is to use our API. There’s no reason to build your own language data-set. There’s only room for us and a couple more canonical companies in our space. This is because the value is in what’s being learned (language practices and patterns) across all our customer activity. Trying this a within a single company won’t give you enough value to compete.

AI, bot, and workflow data: BotChain
The last company I co-founded was a blockchain company. Blockchain can be a good technology for building universal systems of record. BotChain allows companies with bots and workflows to share a common platform to store bot identities, bot actions, and bot decisions. No need for each company to figure out how to operate with each other one by one. Plugging into BotChain will be the universal system of record for bot transactions and identities. (Note: There is a company that will hold BotChain related activities, but the product BotChain is an open-source distributed system.)

Medical data: BlocHealth
The team at BlocHealth in Boston (I’m advising) is putting a universal record of all core medical licensure/credentialing documents on the blockchain.

Examples of future universal systems of record

USR for self-driving vehicles
I’ve had some private conversations with engineers working on the next generation of mobility. In the future (starting to happen now), when one car learns new HD mapping data, the company that made that car will also learn. But is that good enough? With lives on the line, I think that some of that data should be shared for the benefit of all mobility/ logistics. This is where the universal system of record for mapping and self-driving data comes in.

USR for government
I’ve thought that we need to re-think government for a couple decades now. There is a ton of government-related data at many layers of abstraction within government. Some of it’s public and some of it’s private. I’m skeptical that it’s organized and shared optimally to improve government and lower the cost of government. A simple example — no matter what front-end you put on it, a USR should house all the data in cities pertaining to crime, mapping, potholes, etc.

Defense against the looming AI threats

Another benefit of pioneering universal systems of record is that it’s the solution to several looming problems:

The problem of data privacy + AI
Other companies with fewer privacy rights (like China) may move ahead of the US due to their access to their population’s data. This is training data for their AI.

Solution: I propose public (or public/private) universal systems of record that ensure privacy and also allow innovation with the data sets. I’m not sure what this looks like. There’s a lot more thinking to do around rules and encryption. I’m interested in homomorphic encryption for this potential use. If we can get this working we would allow free societies to ensure privacy and also innovate in the area of AI.

The problem of “out of control” AI
People talk a lot about the dangers of sentient AI. I think that will be an issue, but not anytime soon. The real problem will be non-sentient AI that optimizes in ways we didn’t expect. Think of nature. It’s not sentient, but it’s a complex system. And complex systems that impact the real world (like AI will do more and more) are dangerous.

Solution: Systems of universal record can allow us to record, predict, and halt non-optimal decisions (and bad results from those decisions) before they happen.

Next Steps
The best way to know the next steps is to start building. I’m looking forward to see what companies are we need to build to manage common NLP, AI, and IoT data over the next 10 years. Some of them are already here.

Source: On Canonical Companies – Will Murphy – Medium

When we think about uses for artificial intelligence and automation, we often think forward to super-intelligent machines replacing humans. I think that’s ambitious. I think we’ll get there, I just think the next wave isn’t pure AI (artificial intelligence). If you want discuss opportunities, we need to discuss a different AI — and that’s augmented intelligence. Humans will be improved by, but not necessarily replaced by, machines in the next wave. And, that’s going to be great.

I like Kevin Kelly’s saying “The business plans of the next 10,000 startups are easy to forecast: Take X and add AI.”

I would add another one “The best startups of the next decade will use augmented intelligence to create new businesses and markets that didn’t exist before.”

Similar to how exoskeletons will enhance human strength in the physical world across a lot of use cases. Much of the interesting software being built in the next decade will be “exoskeletons for the mind”.

Much of the interesting software being built in the next decade will be “exoskeletons for the mind”.

Garry Kasparov and machines

I was fortunate enough the meet Garry Kasparaov, the chess grandmaster and attend an event he was involved in. He mentioned that while machines could now beat humans at chess, the strongest competitor is the combination of human and machine working together. Sound familiar? Augmented intelligence. It’s here, it’s just not widely distributed, yet.

Moving cognitive load

We will use exoskeletons to move the physical load from our muscles to the exoskeleton — allowing the physical load to move from human to machine. Augmented intelligence will move the cognitive load from human to machine.

I look at augmented intelligence as a mix of technologies to create both intelligent machines, but also insanely simple user experiences. So, my design philosophy centers around three words: Intelligent, Ambient, and Automated.

Augmented Intelligence Framework: Intelligent, Ambient, Automated

The three things I keep in mind as a framework for designing an augmented intelligent system are IAA:

Intelligent: Use artificial intelligence (or any other set of algorithms) to get the job done, and get smarter. The key is to create feedback loops and improve over time based on feedback.

Ambient: Create software that is an intelligent layer over what the human is already trying to accomplish. It should be integrated into the environment as much as possible and assist the human when and where the human needs it. It should assist humans with the lowest amount of user experience friction possible.

Automation: Create software that removes cognitive load from humansand moves the load to machines.

Augmented Communication

At Cognifyd, we’re an augmented intelligence company in the pure sense. Our mission is to move cognitive load from humans to machines, and we’re starting with online communication, something that people spend 1/3 of their time doing at work.

We’ll be helping write business communication for all employees — allowing the signal to cut through the noise and creating more efficiency. It’s time to take one of the things we do a lot of — write communications at work — and augment it. We’ll make it more efficient and more effective.

Humans can communicate. Machines can communicate with humans (see chatbots). The best path to pursue (based on the maturity of the technology) is creating machines that augment humans while they are communicating. Just like chess. It looks like Cognifyd.

We launched our editor today. Sign up on our site for access to our free business proofreading editor.

Stay tuned. We have a lot of surprises in store.

And, pay attention to augmented intelligence. We’re just one of many future augmented intelligence companies.

Sources:
The Big Picture
Infographic City

Improvements in technology may create something better that’s not a block or a chain

Source: In 10 years we won’t have Blockchains – Talla – Medium

I think that in 10 years blockchain technology will improve to a level that calling it “blockchain” will no longer be a useful term. Let’s think about this a moment. Why do we call it a blockchain?

I think the three most interesting parts of the blockchain are:

1. Consensus mechanisms
2. Management mechanisms (meta-consensus of how to manage future improvements to the overall chain)
3. The structure of the blockchain itself (blockchain design, off chain, fundamental architecture)

The last one will impact how large the network can get how fast transaction times will happen, et cetera. And the structure of the blockchain is why we call it a “blockchain.” Let’s break that down.

So, what do we mean by “Blockchain?”

Let’s break out the two terms:

Block: A set of transactions connected together cryptographically into a block. So, a block is a bunch of transactions bundled together.

Chain: The blocks from an endless connected chain connected cryptographically to each other. So, this is a linear chain of blocks.

I think in the future the block will be more of a connected set of transactions (block would not be needed). And, I also think that in the future, the chain won’t be a single directional chain of blocks. I think it will look more like a mesh (or graph).

So, maybe we’ll call it a transaction graph. I’ve also seen the term “tangle” for early similar concepts. Maybe we’ll call it “the graph” much the same way we say “the cloud” today. We will all save information and transactions into this global graph (whether we know it or not).

Maybe we’ll call it “the graph” much the same way we say “the cloud” today.

The closest thing I’ve seen so far that matches my imagined path is a DAG. So, this looks very interesting to me. IOTA and Hashgraph are real-world projects that seem to be on the cutting edge of similar designs, those will be a couple projects watch as well. Let’s look at the DAG concept below.

DAGs (Directed Acyclic Graph)

Some of you may be familiar with DAGs.

From Wikipedia:

In mathematics and computer science, a directed acyclic graph, is a finite directed graph with no directed cycles.

A DAG-model works differently than a blockchain. A common blockchain requires miners to maintain blocks, a DAG wouldn’t need either proof of work or blocks. Transactions are linked together and can confirm previous transactions. In order to submit your next transaction, you need to validate others in the queue. In order to get what you want (your transaction submitted), you have to do some work for others.

Transaction times in this model would be faster the bitcoin (for example) transaction times. Since the miner is technically removed and each transaction validates past transactions, and transaction times can actually go down as more people use the system. And, maybe we could have a non-linear set of of branches that go in several different directions, where many parallel transactions are happening. New models of management would be needed. So, there is a lot of work to do.

But to summarize, this model theoretically gets better as new nodes are added. So it may be an improved model for both fee (gas) and scalability over current blockchain models.

As I mentioned above, in the future we won’t have blockchains. We may have something that does what a blockchain does, only better. There are still challenges to this model, so I don’t know how this will play out. But, I think this design is intriguing, and I’m curious to see how this develops.