10 years ago exactly today, Twitch acquired my company, Curse. In 2004, I was a high school dropout in Paris with no money, no network, and zero business experience. I just loved video games and built a website. We launched Dec 2004. 100,000 visitors in 1 month. 400+ mods uploaded. Working 18 hours a day posting on forums to promote our website and talking to our users. In 2007, I moved to SF to scale it. I made every mistakes you can make, wrong hires, bad product direction, hiring too much, big office space, you name it. By 2009, we were 30 days from bankruptcy. A chance meeting at the W Hotel in SF saved us with a bridge round. We survived, scaled to 1.2B monthly page views, and reached 50M+ users. Team was 130 employees. It started to get stressful to run but I knew and had interviewed almost everyone. I still had to build a huge fucking company. 45M in rev wasn’t going to cut it. I wanted to build a very big product. We built a voice app for gamers. 0 to 5M MAU in 1 year. We felt untouchable. Then Discord launched. Our growth went to 0 and started to turn negative in 6 months. Twitch bought us to join the fight, but Discord won. I spent the next 5 years torturing myself over how I missed the chance to build a generational company. I’ve cured myself from the disease of regret by building @fdotinc and be involved with so many cool teams and stories. My biggest advices are: Work on something you give a shit about You can do anything, be audacious Surround yourself with a great team

every COMPUTER in the World will become an AI agent... I put $300,000 on this to bet my entire career💰🚨 even the legend @Scobleizer is hyped that @MitosisLabs cut AI hallucinations by 98%! AI bubble will burst if we don't solve this problem because AI is lying to you confidently. lot of sauce in this conversation, and Robert asked some really INSIGHTFUL questions. If you are selling AI to businesses, this is your opportunity. TOMORROW we launch Yappy at @Cloudflare HQ during the world's first Agent-Native hackathon. It's going to change how people look at Local Models - making every computer in the world into AI agents. Follow @MitosisLabs for more sauce and some AWESOME AI Alpha to lead the AI race ;)

Ship with an expert by your side @vovy_ai

Mitosis Labs is officially @ycombinator deal partner 🥳 startups in Y Combinator already use @MitosisLabs SDK to Ship products faster while reducing AI cost and hallucinations, to their customers. if you are a YC founder... you can redeem the deal on Bookface. Search for "Mitosis Labs" if you are a YC Partner... - we Submitted our W26 batch application. - startups in YC are already adopting Mitosis Cortex being part of YC will help startups get better support from us and allow us to co-develop & improve our infrastructure through direct feedback from amazing founders and YC Partners. we are also open to providing similar deals to other accelerators and VC firms to support your ecosystem to grow faster by providing a stack that makes AI reliable + accurate, cost efficient, and faster! Comment below to get Access . send DM to work together.

Be there or be

RT @aditya_on_x: It’s time for Demo day ! We are at the Founders, Inc. festival today, where 100 cracked teams will showcase what they hav…

here's the youtube video thumbnail for @shellular_dev demo lol. - multi-agent: claude, codex, pi, opencode & more - use existing subs - persistent terminals - localhost access to see the output - js console to debug setup takes < 3 mins!! demo: piped.video/5OiZvayop0I

elon says AI needs to be maximally "TRUTH SEEKING". i think we've CRACKED it 👇 AI is non deterministic by nature. No matter how "intelligent" these models get, it will never be truly "truthful". @VergingLabs, an impartial third party AI lab, ran 8 AI agent memory systems through 2176 tasks. Look at these impartial benchmarks attached below. @MitosisLabs Cortex got 0 answers wrong. NO HALLUCINATIONS. Not one, out of 272. ✅ 0 fabricated memories. ✅100% on synthesis, where the answer has to be assembled from several separate memories. ✅96% correct overall, 2nd out of 8, beaten only by a hand maintained wiki. Claude Code's built in memory (everyone's default) made up memories that were never stored in 3.5% of its runs, and got only 57% right. why does "truth seeking" even matter? When AI lies, it gives you WRONG answers CONFIDENTLY. For businesses this can be catastrophic. Sending the wrong invoice to your client. AI gone rogue and reaching out to investors with SLOP email. I know this because I have suffered through all of this. One time it randomly started talking about my business to my mom. so what is the ROOT CAUSE? DATA. If the data is bad inside your business, the AI will hallucinate a lot. Doesn't matter if you use Haiku or Grok 4.5. Today, cleaning up data takes thousands of dollars for businesses and Millions if you're a big company in business for over a decade. It is a Trillion $ industry. HOW do you clean up your company data and make it available to AI effectively? Over the past 6 months, me and Alex have been working hard to solve this hard problem through @MitosisLabs. Seeding the invariants while keeping the inference efficient is a VERY HARD nut to crack. But I think we have cracked it. @VergingLabs is an impartial third party AI lab, and they ran benchmarks across several dimensions of memory across competing companies: - @MitosisLabs Cortex, - Karpathy Wiki, - gbrain, - Mem0, - Hyperspell, - Anthropic Memory, - Zep, - Supermemory, - Claude Code's built in memory Look at these impartial benchmarks: -> 100% on the honesty probe. 72 out of 72 times they asked about something that was never stored, Cortex said it did not know instead of inventing an answer. -> 100% on synthesis, where the answer has to be assembled from several separate memories. -> 96% correct overall And this is available TODAY at mitosislabs.ai I am confident that this will help several businesses and consultants selling AI services to businesses see greater profits and ROI.

Does anyone have a large flip disc display we can borrow till August 8th in SF? We're open to buy as well. Any leads are appreciated! Shoot me a DM

Introducing Double. Never apply ever again. ps. whoever invented the job search is an absolute burger. #jobsearch #career #startup #college #SF

"If your engineers say they'll put something in the backlog, they're basically saying fuck you." — @koomen (GP @ycombinator) at the talk at @fdotinc with @hthieblot. He's right. The backlog is where customer feature requests go to die. That's why we're building Trivia, the AI customization layer for enterprise SaaS. Your users describe the features they need and get them instantly, built on your existing APIs. No backlog. No roadmap fights. No engineering time. Shoot me a DM📥 always open to chat if this interests you or check out Trivia at trivia.dev

Excited to share what we have been cooking up at @fdotinc for the past few weeks the era of one size fits all software is over. for 50 years, software came with a deal attached. every customer gets the same screens, the same workflows, the same dashboards. take it or leave it. that deal is not aging well. the average enterprise now wastes $20M a year on software licenses nobody touches. vendors know this. but no team on earth can build a custom version of their product for every single customer. until now :) meet Trivia, the AI app builder that lives inside your product. your customers describe the feature they want in plain english, and it gets built in seconds. on your APIs, with your components, in your design language, with your permissions. it looks like you built it, because in every way that matters, you did. no ticket. no roadmap. no waiting a quarter for a dashboard that takes one sentence to describe. our mission: make software dynamic, shaped by every user who touches it. watch it happen below! if you run a SaaS company and your feature request board haunts you, Book a demo ! trivia.dev

When people ask me why I cancel my Pro Codex subscription:

If your cofounder isn’t following ideal vibe coding ergonomics you’re #ngmi

Let agents build the demo videos for you. Join the waitlist!! trivia.dev

For everyone working on a SaaS products !! - shipping demo's all the time is not fun. especially when your shipping changes every day. Now you can connect your Github Repo and have an AI agent use your own code to understand your product, grab the UI elements and ship a polished demo in just a few minutes ! trivia.dev/ - join the waitlist now, we are onboarding early users ! (this demo was built automatically in a couple of mins ->)

you're on a date. production is on fire. and all you have is your phone. open Shellular: - use Claude Code / Codex / Cursor - run terminals - access localhost - use browser DevTools - test the fix - push to production all from your phone.

We got accepted to @fdotinc Off Season II cohort. people talk about dropping everything and going to SF to build. Its our turn to do it. excited to meet everyone there ! building - trivia.dev

My cofounder @aditya_on_x and I got into @fdotinc's Off Season II cohort. Couldn't be more excited to meet all the cracked people and build Trivia in SF!! See you there 🚀💫