@Propriocetive: I turned down a $4M offer at a $40M valuation several months ago. Came back 4 months later with clear proof of progress…
Summary
A founder shares his experience turning down two acquisition offers (at $40M and $400M valuations) for an AI interpretability startup using a geometric and proprioceptive approach, now with a finished product and open research on Zenodo.
View Cached Full Text
Cached at: 07/05/26, 08:37 PM
I turned down a $4M offer at a $40M valuation several months ago. Came back 4 months later with clear proof of progress — scaled across all major clusters (except B300), significantly beaten competitors in interpretability at their own game. They came back with $40M at a $400M valuation. I turned it down again. We now have a finished product with: • Thousands of unfakeable benchmarks hosted on AWS S3 • Hundreds of GPU hours run on Lambda • Incorporated our superior fiber bundle IP into 6 other companies’ core problems — and outperformed them across the board The geometric + proprioceptive approach is delivering real, measurable advantages. Full patent novelty & value analysis here: https://chatgpt.com/share/6a48341d-d970-83ea-bd2c-f4f26fa52bcf… Additional discussion: https://share.google/aimode/AY878X9 ScloSprCMD @elonmusk @karpathy @sama @GoodfireAI @SakanaAILabs @rohanpaul_ai @xai
https://zenodo.org/records/20808218…
I’m being chased by the few VCs I have researched out to. My goal was always to wrap things up before I made a deal with anyone. I’m in the process of converting from provisional to utility this month and hiring a go to market guy.
I have priority over the next AI architecture. I’ve published 1,500 pages of open research on Zenodo. Anything I don’t patent becomes prior art and the labs die.
What would you do in this position?
Check out this chat
Source: https://chatgpt.com/share/6a48341d-d970-83ea-bd2c-f4f26fa52bcf Get responses tailored to you
Log in to get answers based on saved chats, plus create images and upload files.
Similar Articles
OpenAI Offered Him $2M To Stay Quiet
A former OpenAI insider, Daniel Kokotlo, warns about the risks of AI recursive self-improvement and proposes a safer global plan after leaving OpenAI and forfeiting $2 million in stock options.
How a former DeepMind researcher raised at a $300M pre-seed valuation before launching a product
Former DeepMind researcher Andrew Dai raised a $55 million seed round at a $300 million valuation for his visual AI startup Elorian, discussing fundraising strategy and the importance of visual understanding in AI.
@joanathanlmc: We did $100k in revenue after realizing that our customers hate software. Our first product was an AI accounting tool f…
A founder recounts pivoting from an AI accounting tool to an outsourced AI-powered accounting service, achieving $100k in revenue by focusing on selling outcomes rather than software features.
@neil_xbt: He built a legal AI startup and sold it to Thomson Reuters for $650 million! Then went to Y Combinator and gave away th…
A founder who sold his legal AI startup to Thomson Reuters for $650 million shares a playbook on AI innovation categories, pricing strategies, and the overlooked opportunity of enabling previously impossible tasks at scale.
Jerry Tworek (ex-VP of Research, OpenAI — now CEO, Core Automation) on the two years his idea sat unfunded, and the one sentence that changed it
The article recounts Jerry Tworek’s experience building an AI research idea before it received funding, and how a single encouraging comment from his chief scientist helped move the project forward.