den*

How is a person supposed to tell if they are having a psychotic break or not these days? Everyone else can see the big rabbit listening to Trump talk about bombing Iran, right?

@PeterBell We can swap out shared models for shared contexts and harnesses for the purpose of what I'm wondering about. The cheapest model is the one that's done, as it were, so I wonder how much will be self-spinning, and how much will be shared spinning—to put some spin on it. 😆

I don't really care about the models today (well, a little more after the Claude Max plan fiasco). I find it's all about harnesses, compounding context, memory systems with explicit domain specific knowledge so they don't (for example) remember stuff you're writing to the db) and really thoughtful adversariable quality/verification steps/gates. I find with those you can get the best out of the model you use and you should then use the cheapest model that sufficiently clears the validation gates for any given step in a pipeline/workflow. Perfect world, your LLMs compile every step down to cheap, fast reliable deterministic code. Real world, they minimize the scope/effort of the agentic work and minimize tokens and latency per unit of value.

My Agents Don’t Want to Use Your Laptop I’ve been meaning to write this for months. If you think I’m crazy, make a note and tell me at the end of 2028. I’m curious how this one will age... Peter Bell Apr 05, 2026 Today, the most technical, highest-agency people I know are building agentic systems that genuinely let them do the work of a team. Not “AI-assisted productivity.” Full orchestration: agents that research, draft, schedule, build, review, and ship. One person doing what used to take forty or fifty. Some of these people will stay solo. But some of them are going to want to work with other humans again. Join a company. Amplify their impact beyond what even a great agentic system can do alone. And with the impact of a team of 40, some companies are really going to want to hire them. So let’s play this out. The Offer A technical former CMO has spent the last 18 months tuning their personal agentic org. They decide to take a job at a brand they’re passionate about for a few million a year. HR has a meltdown about pay bands, but the board is pressuring the CEO to turn around revenue, and this person’s track record is undeniable. The deal goes through. Monday Morning The new CMO’s agents send a brief to NewCo’s legal team. It proposes contractual language for limited use of and access to the CMO’s agentic system. The key terms: Ownership and licensing. NewCo owns all deliverables outright. But the CMO’s agents retain a perpetual license to train on the generalized patterns of the engagement (methods, workflows, structural approaches) without retaining any direct record of proprietary company data or PII. Think of it like a consultant who gets smarter with every client but never shares your strategy deck. The DMZ. A technical proposal for a shared context zone: a data environment where the CMO’s agents can access the company information they need to do their job (brand assets, performance data, customer segments, strategic docs) without that data ever touching the CMO’s personal infrastructure. Encrypted at rest, access-logged, with contractual and technical guarantees that the data stays in the zone and gets returned or destroyed at the end of the engagement. Agent-to-agent interfaces. A proposal to stand up a joint working group to negotiate how the two agentic systems talk to each other. Authentication protocols for agent-to-agent communication. Agreed-upon message formats. Rate limits and circuit breakers so neither system can overwhelm the other. Permission scopes so the CMO’s analytics agent can query the data warehouse but can’t touch the billing system. This is the API design problem of the decade, and nobody has solved it yet. Verification and audit. The CMO’s system commits to a continuous audit trail: what data was accessed, what was generated, what was sent where. Not because anyone’s being shady. Because when you have autonomous systems operating across organizational boundaries, “trust me” isn’t a compliance posture. The proposal includes provisions for third-party verification of the data handling commitments by a trusted agentic auditor. Sub-processor commitments. The CMO’s agents use external services (LLM providers, search APIs, analytics platforms). Each one is a sub-processor under the agreement, with the same data handling obligations flowing down. Insurance. Cyber liability, errors and omissions, probably some novel provisions for autonomous system failure modes that no underwriter has priced yet. The CMO’s proposal includes coverage terms because they’ve thought about this. They’ve had to. Meanwhile The IT person rocks up with a laptop and a sticky note with the temporary password to log into Office 365... If this happens, IT infrastructure assumptions are about to get very weird, very fast. This is not a today problem. And it assumes that the advantage individuals can create by staying on the bleeding edge continues to grow. If LLMs are a single S curve, this will probably never happen. Everyone else will catch up and we will all end up using whatever janky agentic system our employers hand to us. But if this is just the start of a series of ever-steepening stacked S curves, and if the innovators on this curve get a durable advantage in climbing the next one and the ones after that (both possibilities, not axioms), this seems like one plausible implication of that trend. I’m not going to ask how your company plans to handle the people who show up with capabilities that don’t fit in your org chart, your security model, or your onboarding checklist. At least not this year :)

@DrBrianKeating What are arots of matters?

A Brief History of Matter

The book has an element of "is this going to be a zero-sum game?" that I was a little surprised they left out of the movie. Since it's Weir, of course it is teamwork, vs backstabbing, but he did have me going a little.

@DrBrianKeating @Arrogance_0024 Why do we take such poor care of our veterans? Why do we listen to chickenhawks? (These are rhetorical.) A little more object permanence from all of us would be nice.

@Arrogance_0024 Combined age of “all this”: 155 years Estimated Value of 1 American Hero: Infinite Any questions?

Lose all this to rescue 1 pilot and call it your greatest military success of all time.

They dumbed it all down when they started knowing people were pregnant before the people did

If you want to design a system from scratch, first you must design the Universe

@knowthyselv @traskjd @agentK So like all things [compute] the answer to if you [will] need it or not is "it depends". This is all common knowledge, but I'm a sucker for stating the obvious sometimes. Maybe you're a prepper, or work for healthcare, etc. All that said, love is really all we need to know!

@knowthyselv @traskjd @agentK Not that everyone needs to know everything, but we do want to be able to fill in when our co-workers win the lotto and leave to hit the beaches, as it were Or our governments gets too hostile, or existing infrastructure starts to suck, or there's viruses, or cyberwar, and so on.

Kind of glad to have begun coding in the 90s. Skipped Kubernetes. Skipped graphql. Skipped microservices. Skipped react unless absolutely necessary. Got mocked a bunch by “senior devs” who started coding in 2010. Didn’t care, preferred shipping value to customers rather than technical wankery that only bloated teams and slowed delivery. Glad those senior devs have now earned their title and realizing it was dumb too. You can’t replace actual experience.

I get the sentiment of course, I'm saying the part that needs work is the wisdom of when to use what. The tech itself does what is needed where you need it, and is generally always getting better.

This guy is broadcast-only. The muting and blocking is especially annoying from people who put Trump in office. (Not that this guy did, just in general.) We cannot block the president. At least have the decency to put someone decent in there. Think of the children!

In a surprise twist, the Epstein files released the attorney general.

@Rtonkins74 @Fat_Electrician The interesting part is how remote service work fits in. Global commerce basically. Otherwise locale-based minimums makes sense. And we've gotta incorporate that somehow.🤔 If only free markets existed. But they don't—unless they're all free, in a "reality is anarchy" sense.

@Fat_Electrician In most cases the market has moved past min wage. Highschoolers make double at rural mcdonalds.

I fucking hate the “minimum wage is for high schoolers” talking point. A. No it isn’t, that’s historically inaccurate. B. Telling a kid to work hard and earn money the “right” way, then turning around and saying he deserves less just because of his age is bullshit. Pay should be based on performance. Nothing else. If you’re against raising the minimum wage, that’s fine. Make a better argument that isn’t belittling kids trying to work hard and earn something for themselves.

"By whom", even, neh?

@Mustang_Scott Some seeming kin—well met! We were talking about censorship, if we were talking about anything. Reddit can be however it wants. I agree it's not particularly "even". You *seem* to be advocating some kind of coerced balancing. By who if not the government? What are ya say'n?

@denstar You seem to be retarded Jon Stewart was saying Reddit is a great example of good social media- im pointing out it’s only good for one side. If Jon Stewart was want a partisan hack he would mention that We weren’t talking about government involvement.

It only takes one person to make an echo chamber, so we're not getting rid of them anytime soon. Which is fine. They come and go. Social content moderation is acceptable, it's the governmental kind that's not. A lot of folk seem to think compelled speech is free speech but no.

I'm hoping with semver and the introspection these LLMs bring, we'll finally have pretty painless semantic versioning for general code. (Like the modelling world has had for a while. But that will be better now too!) Exciting times! 🤘

@therealdanvega I'm not super-sold on "versionless" API. We're persisting data over time, and with versions you'll know what needs to be done to transform a previous format into a new one. There are other ways to track it, and lots depends on the use case, as some people still hard-delete etc.

@therealdanvega Schema-first is great! (Tho you can gen from your JPA entities with a plugin.) I use a generic modelling schema so I can generate most stuff, including graphql schema. I've been digging mapstruct with DTOs, but I do like the idea of skipping DTOs (except for losing mapstruct).

Planned, Recorded, Edited and Published 10 videos in about 6 hours today 🤩 If you're curious about building GraphQL APIs in #Java this is a good resource for getting up and running quickly. Would love your thoughts 👇🏻 youtube.com/playlist?list=…