jonreilly.eth

It’s not about junior vs senior, it’s about “good with AI” vs “not good with AI.”

They could sell this as a service (I would absolutely subscribe).

.@Sonos should build a Clawdbot for the home. Interact via voice (@elevenlabs?), connect to all music services and home control and intelligently integrate it - actually valuable intelligent home unlocked. It could work with existing hardware.

1/ I’ve worked with hundreds of VCs and keep running across the same problem: there’s no simple tool to research rounds and competition. So…we built it. The Venture Codex is a better way to do research before investing and before helping your companies fundraise.

The desired outcome of keeping Bitcoin censorship resistant is shared by all. Just some people are less good at understanding network game theory and untangling critical thinking vs how they would like things to work. bitcoin runs on economic game-theory, not wishful thinking.

I get the appeal. We are told that improving our biology is "unethical", and we want to defend against this charge, and yet we're also often told it's a cardinal rule of public communications that if you're always "on the defensive" you're guaranteed to lose. And so what better solution than to turn around the table and say that those who reject the new century's modernity are the unethical ones, because they are committing the age-old sin of negligence? But ultimately I now think the strategy has far more downsides than upsides. The idea that differences in preferences and values can often be resolved with "live and let live" is a powerful social technology, and we lose a lot if we abandon it. If you publicly telegraph that if you win your fight for acceptance you will immediately follow it up with a fight for imposition, do not be surprised if people decide to oppose your fight for acceptance from day one with a vigor only reserved for resisting imposition. And so if we want the health gains of much stronger biotech this century to come with less opposition, we must put a stake in the ground and make clear that it will all be voluntary. (BTW yes I bite the bullet. I am okay with not mandating vaccination despite the public health risks of such a policy) I call on @noor_siddiqui_ and all friends of radical biotech to join me in making clear that we support all invasive biotech being strictly voluntary, and that we will find other ways to compensate for the downsides of some people not getting certain things at birth or childhood. Milady.

@dharmesh Not certain which is more accurate but HubSpot.eth went from $75K pre-update to $330 post-update.

Thanks to all those that have been helping test my Domain Value Estimation tool (domainvalue .com). Just posted a major new update. Uses more data has more heuristics and some reasonability checks. If you get a chance to try it, please let me know what you think of the results.

Asked the Head of AI at a Fortune 500 how he decides buy vs build, and his framework feels widely applicable: “we buy what accelerates us, and build what differentiates us”

There’s likely too much fear that AI models eat the app layer as they improve. For AI Agents to work, most enterprises will require a bridge between the AI and their specific workflows. It turns out the last mile of making AI Agents work in real, highly variable and hostile environments, is insanely hard. And increasingly it’s the most valuable part of the whole process. That bridge between AI models and enterprise workflows will be a heavy amount of software to connect to different systems, pulling in the right enterprise data, handling security and permissions properly, and having a deep level of context tied to the use case. Then you add in customer support tailored to the use-case, SLAs, liability clauses, tailored sales motions, aligned partnerships for the category, and so on. The list required is quite endless. Every single vertical, and even critical horizontal category, will require a deep amount of expertise to make the AI Agents effective. The big opportunity right now is to identify where these gaps are the widest (between model and the workflow), and fill them in with the appropriate software and expertise. And, even as the models improve - which has previously presented the risk of cannibalization - the focused players can just offer even more value and use-cases to customers. There’s almost no scenario, if you’ve gone after the right market opportunity, where model improvements are a bad thing when building AI Agents.

Hosted a dinner with about 15 IT leaders around the future of AI Agents in the enterprise. Here are a few updates on the state of the world at least at some large enterprises: * For many organizations, the demand from the business for AI is continuing outstrip the ability to implement the technology. This is super relevant because we’re already 2.5 years after the ChatGPT moment and there’s really no slow down in use cases. * AI is causing more department lines to collapse or blur. Companies are finding that teams can now begin to do more of the work of their adjacent functions, which clearly will have some very interesting implications to corporate org structures in the future. * Getting workflows well understood before you add AI Agents to them continues to be a hot topic. If you don’t have a clean process today, it’s very hard to bring automation to that work, so many companies are using AI as an opportunity to bring more discipline to the workflows. * AI Agent interoperability remains a continued focus for IT leaders. It’s obvious that no singular system can handle all the agentic workflows across the enterprise, which means having AI systems talk to each other is still a huge focus. * Training of the next generation workforce remains a huge focus. It’s clear that everyone believes the next generation will work way faster, but there remains some concern on how this new workforce can fully learn the ropes of the business without having to “do the work” now. Overall, momentum remains very strong for AI Agents in the enterprise, and at the same time it’s clear that there will be years of change management ahead to fully deploy agents across the enterprise.

The play right now is to go deep in a vertical and build AI Agents with the context of the critical workflows, domain expertise, specialized instructions, and data of that industry. And build anticipating what’s possible with AI models in a year from now, not just today.

@morganlinton Grabbed ctoconvo.com - if you like it lmk and it’s yours, otherwise I’ll let it expire

Thinking of names for my podcast, and my first plan was to keep it simple and go with CTOsPodcast(.)com. But I'm wondering if maybe that's too boring? 🥱 So here's some other ideas, let me know if any of these hit home more? 🎙️ CTO Pulse 🎙️CTOs Unplugged (could do fun artwork based on the Nirvana Unplugged album) 🎙️ CTO Stories Or is it none of the above, back to the drawing board? Of maybe you like CTOs Podcast, let me know what you think 🧠 Unless you're already using @cluely in which case you can't tell me what you think because you no longer do.

+1 for "context engineering" over "prompt engineering". People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window with just the right information for the next step. Science because doing this right involves task descriptions and explanations, few shot examples, RAG, related (possibly multimodal) data, tools, state and history, compacting... Too little or of the wrong form and the LLM doesn't have the right context for optimal performance. Too much or too irrelevant and the LLM costs might go up and performance might come down. Doing this well is highly non-trivial. And art because of the guiding intuition around LLM psychology of people spirits. On top of context engineering itself, an LLM app has to: - break up problems just right into control flows - pack the context windows just right - dispatch calls to LLMs of the right kind and capability - handle generation-verification UIUX flows - a lot more - guardrails, security, evals, parallelism, prefetching, ... So context engineering is just one small piece of an emerging thick layer of non-trivial software that coordinates individual LLM calls (and a lot more) into full LLM apps. The term "ChatGPT wrapper" is tired and really, really wrong.

Calling c the "speed of light" completely misses the point. Rather, c is the "spacetime exchange rate": how many units of space you can exchange for one unit of time. In actuality, everything travels at the "speed of light", just not necessarily through space alone... (1/4)

Why I support privacy: vitalik.eth.limo/general/2025/0…

We're excited to share our strategic partnership with Horizon Media, marking a new era of AI-driven advertising. With Akkio’s AI marketing analytics platform, Horizon Media will be able to analyze campaigns at scale—faster and more efficiently than ever before. Read more here: akkio.com/horizon-media-… #HorizonMedia #Akkio #AI #AIMarketing #AIAnalytics #AIAdvertising #MediaPlanning @horizonmedia

ChatGPT currently down.

AI isn’t here to replace data analysts; it’s here to unlock their full potential and dramatically scale their work. I share my thoughts in my latest on @Dataversity #DataAnalysts #Analytics #Data #GenerativeBI #GenBI #AI #ML

Generative BI is the intersection of Generative AI and business data. I’ve written about our vision for Generative BI below. I’d welcome your thoughts about how you foresee GenBI transforming your small business. #GenerativeBusinessIntelligence #GenBI #AI #ML #GenAI #DataAnalytics

The future of working with data is #AI-powered. Akkio just raised $15M to make generative AI accessible for any business. This is the start of a profound movement to redefine work in the era of AI. Read more: hubs.la/Q01ZCmTn0 How @AkkioHQ is leading the charge 👇