Simon

@itsolelehmann Can you share any evidence to back this claim: “MRI-level detail”

i absolutely love "pure" startups like midjourney > took zero outside money > fully bootstrapped > profitable since basically week one > doing around $500m a year > with a team of like 150 people. which means when the CEO david holz decides he wants to build a sci-fi full-body scanner... one that lowers you into a pool of water, maps your insides with MRI-level detail at ~100x the speed of an MRI, with no radiation and no giant magnet, there's nobody he has to convince. he just builds it. because there's no board or VC asking how a medical spa fits the image-gen roadmap or what the TAM looks like. founder control is the cheat code. it's what lets someone make the cool, risky, slightly absurd bet everyone else is too scared to even propose. more companies like this please.

I built an internal system at Centr that takes Figma designs to production. It’s easy… - designs done in Figma - Figma MCP - agent + human QA loop - Github/Vercel pipeline There’s a separate flow using Sanity CMS and our internal CMS to conjure pages using templated components

In the great words of @pmarca: interesting Probably the last pages I do the ‘old way’ try.centr.com/i-could-never-… try.centr.com/get-off-the-si… try.centr.com/this-is-all-it…

@geeGesz My mate in US is watching fine

@sihulton Mines not working either

@karpathy @bcherny *commando rolls outta bed*

This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November), peaking especially for long problem-solving sessions on very difficult problems. You can give it a lot more ambitious tasks than what you're used to, the model "gets it" and it will just go, and it's never felt this tempting to stop looking at the code at all (but don't do this in prod!). The model still has quirks that people will run into and the safeguards are configured to be a little too trigger happy for launch, which can hopefully be tuned over time. I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing substantially. You can ask for anything - explainers, visualizers, dashboards, bespoke single-use apps (e.g. a full wandb that is hyper-specific just for your project), you can 10X your test suite, auto-optimize code, run giant research projects with custom HTML for the results, anything! "Free your mind" (Matrix ref). Really looking forward to all the things people build!

@WallStreetApes Heaps of people with big followings are publicly anti Pratt — dj_dave has 250k followers for example

The election has just been called for Los Angeles Mayor Karen Bass, she is advancing to the General Election Without a single ounce of support or any positive comments on any social media platforms, she takes the win California elections are so obviously rigged it’s insulting

@teddypowday CTO / COO combined is where it’s at Teddy and must also be The AI Guy

Your CEO should be strong. Your CTO should be wise. Your COO should be wicked, cunning, of mysterious origins, fluent in the dark arts, blurry in pictures,

Thanks Paul, nice to see someone else who actually knows what they’re talking about amidst this sea of ignorant talking heads with no experience, that have somehow amassed followings… I fully dropped out of using Reddit because of the level of ignorance; unfortunately pervasive here too.

Sorry, but in enterprise AI we’re building stunningly usefully fully automated apps, tools & human-in-the loop systems every day, that were impossible 1-3 years ago. Just RAG itself is stunningly useful in business. But the decision making apps and internal tools too. We spin up tools with UIs in a matter of minutes now. It’s not remotely just summarize this PDF. Nobody even bothers to try to measure the ROI. Your boss knows the ROI. It’d be like asking if there’s ROI in using computers or pen & paper.

we are being gaslit about AI on a societal level. Everybody is vibe coding but I haven’t seen one useful thing get produced. Everybody has agents doing something but nothing useful is getting done. Cool you had AI summarize a PDF and make a template. Nice

@jerrydotxyz 🫡🦘

Nothing more appetising to an American venture capitalist right now than an Australian accent

@rileybrown Why the app. Chrome DevTools in VS Code running Claude Code. What do you think it can’t do already?

As soon as you shove a full browser into a desktop app that works seamlessly with the agent that is game over for separate apps. You don’t need “Claude Design” or “Google Stitch”. If the user asks for a design, the design app opens in the Browser and the generated design shows up there. It seems like a foreign idea to most people now but I think people will pick it up really quickly.

For those who are infinitely curious, there is no greater tool that could’ve fallen into our laps than AI

When Claude Code first came out I was so hooked I took my laptop shopping with me

Here are 3 things I’ve built recently as a non-engineer (product -to- Director AI). These are for my friend’s company not where I work. 1. his Sales guys go to heaps of trade shows this is like an intel app — it uses Hubspot + scrapes to create dossiers on everyone attending 2. a Vercel/Slack scoring pipeline pulls ~50 datapoints to prioritise leads (Hubspot, Quickbooks, Shopify) 3. a video player for a product they sell which is a TV inside an MRI room — Electron app on a tablet via HDMI to the TV — plays 5min clips that display on the TV I’m a bit ahead because of my technical product background, but domain experts will be REQUIRED to express their value as software. I have zero doubt.

@SatoshiWolf Mandatory street parties

If you're in charge of Australia with the ability to pass any policy how would you fix the once greatest country in the world?

Agreed, and it’s a challenging problem to solve!

This is what I’ve been doing for the last year

Hi @levie we could have a great conversation about this. We have a complex org with legacy systems; consolidating to Databricks, Triple whale, Netsuite and other systems. We have a monolith CMS that’s archaic and unfortunately deeply embedded. Instead of migrating we’ve built an abstraction system on top of it which uses Playwright scripts to control it! Bit of a monster but that abstraction is super agent friendly and has enabled huge efficiencies and quality improvements.

This is effectively the #1 problem for AI agents in the enterprise. As we go from agentic coding (where a large amount of context is in the code base, and users are technical enough to get the rest to the agent easily) to a world of knowledge work agents, the context problem becomes much more acute. We see this every day with customers at Box. For existing digital knowledge, it’s often fragmented across legacy systems or environments that don’t play nice with agents, and have access controls that don’t map to the real work that needs to be done, which become a huge hurdle for getting agents the context they need. This has to all get moved to modern, secure cloud environments. But also, companies often haven’t captured and digitized some of the critical context that agents need to work with. Decisions, processes, and workflows often live in people’s heads and tribal knowledge that need to get turned into unstructured data for agents. This is actually one of the biggest points of leverage for applied AI companies, because they can work to specialize in getting agents exactly the information and domain expertise they need. But it’s also one of the reasons why FDEs and new system integrator plays will also work so well right now. The companies that figure this out will be able to get the most out of AI going forward.

The future belongs to people who combine human creativity with AI execution.

@rzdchxrry Chess

Genuinely what is the purpose of butt plugs What the hell do you need them for