tag:dlvr.it,2018-03-23:ac6861e9e2d249fc92b2895b5ff7a2e0 2026-06-07T23:14:36-07:00 tag:dlvr.it,2026-06-07:cd4b24d8d7142a5e9acef91b7ddd96f1 2026-06-07T23:14:36-07:00
"All of that's true, except for the reality we don't quite understand the whole recursive self-improvement loop in humans. Recursive self-improvement requires intuition and the ability to learn from missed opportunities and errors. We've created models that are great wordsmiths (sometimes) and coders. There are limits to the AI vs. human equation. After way too many words about how Claude is great at everything and running Anthropic, the company gets to the punchline. Anthropic lays out a few scenarios and then drops this gem. " ]]> tag:dlvr.it,2026-06-07:4ee8db35a162a55ad1179cb4e910dbc8 2026-06-07T23:14:36-07:00
"What a SOC needs is a way to walk each recovery path through an audit grid with its AI build team before the next renewal closes. The AI Authority Audit Grid at the end of this article maps every authentication write a support agent can make on the recovery path, what Meta's incident proved about each one, why it stays dark to the SOC, and the control that closes it." ]]> tag:dlvr.it,2026-06-07:5f9920925945cd51f41373b535084590 2026-06-07T23:14:36-07:00
"Enterprise technology buyers are currently standing on the edge of a potentially incredibly expensive mistake: buying into generative AI hype cycles without auditing their underlying data architecture. To successfully navigate this transition without incinerating corporate capital, buyers must anchor their strategy in realities rather than vendor press releases." ]]> tag:dlvr.it,2026-06-07:82682bafe6009758cab72e3b3e01f481 2026-06-07T23:14:36-07:00
"Ian Goldin, a threat researcher at Lumen’s Black Lotus Labs, said we’re entering unchartered security territory as more large online platforms start allowing AI chatbots to handle sensitive account recovery requests. Just like human customer support employees can be social engineered into providing unauthorized access to someone’s account, AI bots are equally eager to help and vulnerable to persuasion and trickery, he said." ]]> tag:dlvr.it,2026-06-07:e89d4997b95488ddae8fb73861c5a5d4 2026-06-07T23:14:36-07:00
"Have a “cheap” gate keeper : Route the incoming questions to the appropriate component to answer. “What is my account balance” doesn’t need even an LLM – it can be answered by an API call or a DB lookup. Only route complex reasoning tasks to frontier models. Another elegant solution that is often missed is semantic caching – where a recently answered similar question can help reduce the cost of answering the new question to nearly zero. " ]]> tag:dlvr.it,2026-06-07:6f5754681356e6c2aca3e15fcedc9042 2026-06-07T23:14:35-07:00
"It remains to be seen whether or not the industry can adapt to a definition of open that depends on a sliding scale rather a fixed yes/no. But it also doesn’t have a choice. Two years of development and discussion and two years of living with a proposed definition have gotten us no closer to an industry consensus. Subtly, however, what the G7 nations have done with this document intentionally or unintentionally is to both acknowledge that fact, make it irrelevant and implicitly propose their alternative. The challenge for any single definition of open source AI is that it is not possible to please both definition purists and definition pragmatists. The former point out that any definition that allows for any omission of training data is effectively granting the term open source to a project that cannot ever be independently replicated. Which is legitimate. The latter, on the other hand, point to issues with datasets ranging from the byzantine nature of data licensing to the sheer impracticality of the size of these datasets. Which are also legitimate. You can please one of these groups about an open source definition, but not both." ]]> tag:dlvr.it,2026-06-07:cdef32129af0c20ff8ab17d38823b560 2026-06-07T18:58:13-07:00
"But here's what the same data also shows: A meaningful group of companies is breaking the pattern. They are realizing the savings they targeted, deploying agents with genuine confidence, and funding the next wave from returns that actually materialized. They didn't get there by finding better technology or bigger budgets. They got there by treating data access, governance, and process redesign as CEO-level problems rather than IT problems. The gap between these companies and everyone else is widening. Understanding what separates them starts with three uncomfortable truths." ]]> tag:dlvr.it,2026-06-07:7e62fb53acd65a2c8162abbec6dc31ab 2026-06-07T18:58:13-07:00
Enterprise hits and misses - does the agentic enterprise have a friction problem, and is MCP ready for prime time? https://ift.tt/gNlGJUE This week - agentic enterprises aim to reduce friction - but is that a problematic goal? Bring on the debate. Also: vendors incorporated MCP with urgency, but what should customers be asking? Do CEOs really need AI avatars? As always, your whiffs. via Jon Reed – diginomica https://diginomica.com June 1, 2026 at 04:34PM ]]> tag:dlvr.it,2026-05-31:2f67a355efd2926da53181849764258a 2026-05-31T15:44:15-07:00
"No vendor should own control planes for identity, data storage, model routing, workflow orchestration, integration, observability and governance unless there's a business reason. Avoid farming out your universal AI layer to platforms that belong to your existing SaaS powerhouses or hyperscale cloud providers. It just results in lock-in. Your current roster of big vendors should be participants in the enterprise architecture not the architecture. They all want you to use them as the enterprise architecture to manage their inhouse AI agents as well as third-party versions." ]]> tag:dlvr.it,2026-05-31:a498cfd8cd77c186a7a9ced1eeff1168 2026-05-31T15:44:15-07:00
"Yes, it's that time. As the spring event season winds down, we ramp up our month in review. What did vendors get right? What did they utterly miss? What underrated stories do buyers need to track? Where does AI fall short - and where is it getting results? We'll have Brian's infamous slide deck, and, as always, field all your live questions and comments in real time. We may even bring you onto the show if you bring your A game, so bring your sharpiest and snarkiest commentary and let's see if we can actually draw out some key lessons from the enterprise year to date." ]]> tag:dlvr.it,2026-05-31:c755710410736263b1979c12fd7d2660 2026-05-31T15:44:15-07:00
"To meet the CRA’s standard of due diligence, organizations must provide streamlined, standardized evidence that their products are built correctly and maintained securely. Security and compliance leaders must immediately build a readiness plan. Consider this: Auditing and gating every security practice — from thousands of daily commits to production deployments and post-deployment monitoring — is a massive coordination effort. Incorporating new compliance requirements into daily workflows is highly time-intensive, especially as AI radically accelerates development. " ]]> tag:dlvr.it,2026-05-31:5faae73e3b6191cc37dec57cc118887f 2026-05-31T15:44:14-07:00
"For enterprise customers, that gap is worth probing. The MCP integration your vendor announced may not have the guardrails you’d assume. Poorly scoped implementations may expose more data than intended. And the absence of clear terms around third-party AI access creates uncertainty that, left unaddressed, becomes your problem as much as theirs." ]]> tag:dlvr.it,2026-05-31:a936710d798632595d9acb364393b3e1 2026-05-31T15:44:14-07:00
"The Verizon 2026 Data Breach Investigations Report, also released in May, confirmed that credential theft dropped to 13% of breach initial access vectors. Vulnerability exploitation took the top position at 31%, displacing what Verizon called the longtime leading initial-access category. That's three independent sources, same structural finding. MFA protects password-based authentication, but the attacks dominating financial services increasingly bypass password theft through resets, token grants, and exploitation. The MFA Bypass Exposure Audit Grid at the end of this article maps all five confirmed attack surfaces from the CrowdStrike, FBI, and Verizon reports, what MFA misses on each one, and the specific fix for Monday morning." ]]> tag:dlvr.it,2026-05-31:c9b171bc44a00d18185dc077983238db 2026-05-31T15:44:14-07:00
]]> tag:dlvr.it,2026-05-31:925c01705df6cb01f1746ddb03feb1f0 2026-05-31T15:44:14-07:00
"Do not accept Palantir as a bundled line item. Require a detailed breakdown of the tools being deployed, the resources committed, and the measurable impact expected from each. If your SI cannot articulate what Palantir is doing, when, and with what outcome, that is itself a risk signal. Hold Both the SI and Palantir Commercially Accountable SIs are currently voicing confidence while stopping short of contractual commitment to outcomes. Close that gap. Ensure your agreements hold both the SI and Palantir jointly accountable for tool performance, timeline adherence, and project outcomes. If neither party will stand behind the promise commercially, that should weigh directly on your decision to proceed." ]]> tag:dlvr.it,2026-05-31:0d802d219235e186887732627feab683 2026-05-31T15:44:14-07:00
"EM needs more EQ than IQ. Great engineers are a diverse lot – some need to be left alone, some need active coaching, and some will burn out if you don’t keep checking in to course correct. It’s a tiring job and you need to thrive in that environment They need to be communication wizards – negotiating with product managers to translate the roadmap into things that can be coded, while also explaining to finance why you need budget to tackle technical debt. I have always felt that it’s easier for an EM to learn basic business than to teach tech to other stakeholders. I have often felt my MBA was wasted effort but grudgingly I will accept that it helped me explain tech better to others in a language they can follow. " ]]> tag:dlvr.it,2026-05-31:c4649eb80312d725d1516a75b6ccfda5 2026-05-31T14:36:13-07:00
"AI accelerates creation, but it introduces new layers of review, fact-checking and correction," said Evan Reiss, senior vice president of marketing at Foxit Software. "Work is not disappearing. It is being redistributed. The future of document intelligence will not be defined by more AI features but by better integration, reduced validation burden and systems that augment human capability rather than replace it." Trust Remains the Primary Barrier Across all sectors and roles, the biggest blockers to deeper AI adoption are human-centric concerns. Data privacy and security concerns were cited by 36% of respondents, trust in AI output by 34%, and response accuracy by 25%. These concerns can outweigh the demand for new features or a clearer return on investment.]]> tag:dlvr.it,2026-05-31:9df3a95da37f34b197baf5497aefd554 2026-05-31T14:36:13-07:00
Enterprise hits and misses - Uber and Microsoft put the brakes on the productivity of AI tokenomics, while Google bites the data hand that feeds its AI search https://ift.tt/GiJgpuF This week - as event season peaks again (and again), we've got your field views. Does tokenomics stack up against productivity? Microsoft and Uber put on the brakes, and I give my take. Google changes up AI search - what does it mean for sites like diginomica? As always, your whiffs. via Jon Reed – diginomica https://diginomica.com May 26, 2026 at 03:09PM ]]> tag:dlvr.it,2026-05-25:db3cc5053e0a0bc561c123e0a9254582 2026-05-25T20:17:36-07:00
"Establish AI-specific incident management. Clear ownership, escalation paths, and runbooks designed for AI failure modes—not legacy ones retrofitted to fit. Treat incident management as a cross-functional discipline, not an IT afterthought. Define AI’s operational boundaries. Use the three-tier model above to specify which tasks AI can automate safely and which require human judgment. This is the foundation for faster operations and sustainable AI value. Invest in AI observability. Traditional monitoring tools weren’t built to catch model degradation or agents making unintended decisions. Specialized LLMOps solutions can spot these warning signs before they reach customers. Build continuous learning into the process. Every incident should feed back into runbooks, automation rules, and escalation logic. This is how organizations reduce risk over time instead of just reacting to it." ]]> tag:dlvr.it,2026-05-25:31d4a95ad38ea84aac8eb85e4de22491 2026-05-25T20:17:36-07:00
"Mini Shai-Hulud does not stop at CI secrets. Datadog Security Labs documented that the payload reads ~/.claude.json and exfiltrates it. It scans for 1Password and Bitwarden vaults, Kubernetes service accounts, cloud provider tokens, and shell history files where developers paste API keys. StepSecurity's deobfuscation confirmed that Mini Shai-Hulud harvests Claude and Kiro MCP server configurations, which store API keys and auth tokens for external services. For developers using AI coding agents, the worm already knows where their credentials live." ]]>