The future of building startups with AI:. - No-code website.- Generate demand before you build.- Community-driven go-to-market.- AI as an intern, peer, or advisor, depending on the task.- AI coding co-pilot: unit tests, simple features, debugging.- AI marketing co-pilot:.

RT @thealexker: Dropping gpt-oss-swarm🧠, parallel reasoning to reduce hallucinations. Generate 10 responses in parallel with asyncio → scor….

try it out here:

Dropping gpt-oss-swarm🧠, parallel reasoning to reduce hallucinations. Generate 10 responses in parallel with asyncio → score using LLM judge → synthesize the best parts. More complete reasoning that catches edge cases single runs miss. Great for creativity and getting multiple

is RLHF being replaced? the US DeepSeek just dropped, an opensource model competitive with Claude Opus. cogito v1 is trained using iterated distillation and amplification (IDA).> step 1: higher intelligence capabilities via subroutines (specialization).> step 2 distilling.

underrated trait to be successful: being self-aware and honestly indexing your skills, then shoring up weakness or building on strengths.

was told at yc ai sus by mit phds that they have an in-house recruiter that helps them negotiate multimillion contracts, very wise niche for folks working in talent.

RT @thealexker: @im_roy_lee @cluely distribution > software, vibes and brand is the moat. working on releasing an open sourced version of v….

RT @thealexker: You need to be ai maxxing. You need to tap into incomplete GitHub repositories, gather know-how from builders, implement fr….

@_mohansolo @ycombinator thank you @garrytan and team for the amazing conference:) met so many amazing builders/future collaborators.

If AI writes 90% of your code, what will developers actually do?. The CEO of Windsurf (acquired by OpenAI for $3B), Varun Mohan @_mohansolo, gave a talk at @ycombinator AI Startup School on how to build successful AI companies and the future of coding. My top takeaways:.>

RT @thealexker: The real edge of Grok is that X provides a continuous stream of RLHF. 500 million Tweets a day means bookmarks, likes, and….

the oversaturation of 'ai for x' is like the 'uber for x' movement a few years ago. duplication works well when it's problem focused, or better yet, *need* focused. rather than ai as the offering, if ai is used to augment the offering, specifically using its reasoning abilities,.

yet truly great aesthetics and judgement might still be in short supply. but now at least the bottleneck of creation is removed.

chatgpt was the watershed moment for language. sora might that for art. it's the new renaissance and we are living through it. anyone can be an artist and can define the bar for aesthetics.

RT @thealexker: The only 5 links you need to understand the Transformer:.(1) Let's build GPT: from scratch, in code….

spotify wrapped this spotify wrapped that how about you get wrapped in someone’s arms instead.

@ilyasut Bonus advice for those building AI products: .1. Use special data. 2. Imagine where things will be in a few years and how this will affect the product's assumptions, e.g., if hallucination is fixed and output is reliable, can you build towards products that take this into.

Can we predict the future capabilities of AI? . @ilyasut, co-founder and chief scientist of OpenAI, on the scaling law and an interesting area of research:. (1) The scaling law relates the input of the neural network to simple-to-evaluate performance measures like next-word

Will AI be the snake that eats its own tail? . We desperately need to escape the ouroboros of training data. This is a compounding problem as AI-generated data on the internet continues to outpace those generated by humans. Model collapse is inevitable when we lose data from

Another one bites the dust. 🌬️. Olive AI (healthcare), Convoy (logistics), WeWork (real estate). Now Veev. Growth-stage startups that overcapitalized in 2021 are facing a reckoning. Growing without finding real PMF burns you in a recession—more power to cashflow positive and