real_marcfawzi
Re:Mark Zuckerberg freezes AI hiring amid bubble fears
on Thursday August 21, 2025 @04:13PM (#44977486)
all news is being manipulated to short stock and make a ton of money shorting stock, based on perceived bad news
Re:A Spellchecker Used to Be a Major Feat of Software Engineering
on Tuesday August 12, 2025 @12:39PM (#44878729)
i can't help but make funny connection between "warm" making them less reliable/consistent and the temperature setting....
We Fixed AI's Broken Promise
feels like O3 but faster (in quick answer mode), not any smarter .. thinking mode takes forever and the results are mediocre
Re:AI ethics is being narrowed on purpose – just like privacy was
on Thursday August 07, 2025 @10:44AM (#44825147)
Thanks for your curiosity and interest. I need to update the Roadmap because it omits the fact that we already trained the model on synthetic data generated from 100% AI enactments (no humans involved) which baked-in the moral critical thinking framework that we had landed on (through lots of thinking and feedback from people over the last two years.) What Phase 4 is about is doing it with human generated data, where our community members participate in the enactments, but to do that we have to have enough users to generate sufficient data for post training. That way it'll be "community-led" training. Phase 3 is simply taking the responsive web app and putting it inside a WebkitView container, but no point in it until we have funding to scale on the backend.
Happy to jump on zoom and discuss the details. We have raised around $100K in "micro investments" via SAFE at a relatively low valuation, and some of the micro-investors are also collaborators on the technical and/or vision/community building side. Myself and Syd have put in a lot more than that into it, but that was the cost of all the experimenting and learning. Microsoft also funded us with Azure credits for $25K, which we used to partially cover the cost of the initial post training.
Happy to chat offline: marc.fawzi on gmail