Ignat

@omarvvvr Out of speed

What do you think kills most startups today? -out of money -out of motivation

@markproduct Depends on the founder Getting beyond localhost is still hard for many of them 😄 More importantly, AI makes building cheaper and faster, but it doesn’t turn everyone into a builder

If a founder can build an MVP in 3 days with AI for $200... Why would they pay a developer and designer $25,000 and wait 2 months? What's the argument?

Wearable at first, implanted later. And I think it already started. We just haven’t gotten the iPhone version of it yet

@ceozaitsev Execution used to be the limit. Now it’s attention. The machine never idles, so neither do you. I used to burn out doing it all myself. Now I just move too fast to burn out… adhd heaven 😀

@ignat_en Work feels differently with CC agents running for you - they don't stop when you do, and somehow that makes you not want to stop either

Working with Claude Code has shown me something I didn't expect – work can be addictive. Recently I read a post by Nikunj Kothari from FPV Ventures: he wrote that he can't watch a movie anymore without a laptop open next to him, and that reading a book now feels like a luxury he can't afford. I could relate to that a little too well. Since March 30, I haven't had a single day without working on something. Sometimes it's my job. Sometimes it's research. Sometimes it's one of my startups. Usually it's all three at once. And maybe... let's test one more idea. Go to eat? I leave a task running. Time to sleep? Let me start a few agents and loops, so they keep writing code through the night. By morning there's already work waiting for review. Watching something? Of course the laptop stays open (the last season of The Boys is shit anyway). And the strangest part is that none of it is forced. Nobody is making me do this. Most of the time I just want to see what the agents bring back: what got built, what got discovered, what moved forward while I was away. And somehow that turns into 70–80 hours of work every week. While having a full-time job and raising a kid. I'm not sure this is healthy. But it definitely feels like a new way of working.

@markproduct Cook

forget Ai replacing humans. what's one thing you can do that Claude can't?

@evolutionplusai The interesting part of OKF isn’t the format itself. But if it gets adopted broadly, we’ll finally have a common way to move knowledge between tools instead of rebuilding integrations every time

@ignat_en Format was never the hard part. Keeping the context true after the third person edits it is.

Google just shipped the Open Knowledge Format. It's good, but it skips the hard part. Markdown plus a little YAML frontmatter, shippable as a tarball, readable in any editor, indexable by any tool. If your AI context is trapped in scattered wikis and code comments, OKF gives it a clean, portable shape. I agree with the diagnosis. It's the problem I've been working on for the last few months. Making knowledge portable is an important step forward. But here's what the announcement leaves out. The format is the easy 80%, and Google just gave it a common language. The hard 20% is everything a static file can't represent: 1) Who's allowed to see which fact. A markdown file in a tarball doesn't know the comp doc is off-limits to the intern. A company brain has to enforce that inside the query itself, or the AI will happily quote a source the asker was never allowed to open. 2) When a fact stopped being true. OKF has a timestamp. A real decision has a lifecycle: it took effect in January, got superseded in March, and the thing that replaced it points back. Ask about a year-old decision and you should learn what changed, not get a stale file with a confident date on it. 3) Whether the answer is proven or guessed. A wiki hands an LLM context and hopes. Grounding means every claim is pinned to the exact sentence it came from, with a page and a source you can open. Readable-by-an-LLM is not the same as verifiable. The format makes knowledge portable. That's a really important step forward. But it doesn't make it permissioned, dated, or provable. That's the actual company brain, and that's where the real work is. We're building it at Combra. If this is the problem you're living with, the waitlist is open.

@GetBrainpage Exactly. Documents are evidence, not truth. There has to be a decision layer above the documents, because decisions are made by people, not files.

@ignat_en the part that broke us worst wasn't access or lifecycle — it was conflict. three docs answer the same question differently and a flat file has no notion of which one wins. are you encoding authority/precedence somewhere, or punting that to the query layer too?

@ThePeterMick I can do this all day combra.ai

drop your startup link

@GoogleCloudTech Really like this We've been moving in the same direction at Combra and are about 90% of the way there already If OKF catches on as a standard, that's great for everyone, and we'll be ready to plug straight in

Introducing the Open Knowledge Format (OKF), an open specification that formalizes the LLM-wiki pattern into a portable, interoperable format. AI is only as smart as the context we give it. As we build more advanced, agentic AI systems, they need accurate metadata and context to be useful. But in most organizations, that context is locked inside fragmented data catalogs, isolated wikis, scattered code comments, or the minds of senior engineers. Every time a new AI agent is built, teams are forced to solve the exact same context-assembly problem from scratch. To solve this, we've announced OKF, a vendor-neutral, open specification that formalizes the "LLM-wiki pattern" into a portable, interoperable format. It provides a standardized way to represent the enterprise knowledge that modern AI systems rely on. — Just markdown: readable in any editor, renderable on GitHub, indexable by any search tool — Just files: shippable as a tarball, hostable in any git repo, mountable on any filesystem — Just YAML frontmatter: for the small set of structured fields that need to be queryable: type, title, description, resource, tags, and timestamp We’ve also shipped reference implementations to help you hit the ground running, including an enrichment agent for BigQuery, a static HTML visualizer, and live sample bundles on @github → goo.gle/4uGvAEe ➕ Knowledge Catalog can now natively ingest OKF! Stop reinventing data models and building bespoke integrations for every new AI tool. Here's more about how OKF works → goo.gle/4uGvBbg

@looksmaxxvc you seem emotionally invested in this

@ignat_en You are not based you are a subhuman.

So based. I used to think Oleg Tinkov was cringe. A guy who swore constantly, talked about how his company would fuck everyone else over, and generally behaved however he wanted. But then the war started. Out of all major Russian businessmen, he was one of the very few who publicly spoke the truth. Because of that, he lost the company he had built — a bank that was once worth around $9B. But he started over from scratch, and just 4 years later he built another bank, Plata, that's already worth more than $5B. On the other side of the world! And the same goes for Elon Musk. Sure, he makes a lot of controversial statements. Yes, some of them are wrong. But he's also an exceptionally talented entrepreneur who keeps proving it through what he builds and through his position on the Forbes list. It's not scary to be cringe. What is really scary is being nobody and achieving nothing with your life.

@SlackHQ @trq212 Finally! Thanks!

@trq212 The gap between 'AI generated something useful' and 'a coworker actually opens it' is mostly a rendering problem.

Slack now renders HTML attachments instead of just showing it as text 😭🙏

I recently listened to a podcast about modern education and how we should be teaching kids. At one point, they were discussing the social media bans in Australia, and an 11-year-old girl said: "If my grandmother lives on the other side of a forest full of wolves, teach me how to walk through the forest. Don't forbid me from visiting my grandmother." Remarkable that even small children seem to understand this better than many modern politicians.

Two years in what I still think is the best country in the world. With all my projects, I almost missed the date entirely. Funny how quickly even a place this extraordinary becomes part of everyday life.

@dflieb If your company brain makes you search, you’re doing it wrong. Join Combra’s waitlist to see how it should really work in the AI era 🙃

It's so nice having our YC company brain have access to slack, so I can use my agent to....search slack.

@16vchq combra.ai

drop ur startup link

Give your company a memory that doesn't leave when people do. Join our waitlist. combra.ai

It's a great feeling when industry leaders put your own vision into words: the ability to switch out a "generalist" model without losing the "company veteran" expertise. Everyone is racing to get the best model. I think that's the wrong race. The model is turning into a commodity, rented by the month and swapped whenever a better one ships. What is not a commodity is what your company has actually figured out: the decisions, the "we already tried that, here's why it didn't work," the rule that quietly replaced an older one. The problem is this knowledge has never lived in one place. Customer conversations sit in one tool, product decisions in another, the analytics in a dashboard, and the reasoning in a recording nobody reopens. The threads between them exist almost nowhere except inside a few people's heads. And that's what makes those people so valuable. They don't just know facts. They remember that the feature you're debating today traces back to a complaint from six months ago, or suspect that yesterday's drop in conversion is the pricing experiment someone ran weeks ago. What excites me about AI is that those connections no longer have to live in a few people's minds. The dots a veteran connects instinctively, a model can now connect for everyone, across everything the company has decided and shipped. You can always offload the work. The learning is the part that should compound, yet it's exactly the part we keep losing. A new hire shouldn't need six months to figure out how the company reasons. An agent shouldn't start from zero every time a new model ships. What the company has learned shouldn't vanish because the person who remembered it left. The model underneath will change. The understanding on top shouldn't. That's exactly what we're building at Combra.