Sam Hajj

Founder, CEO @ Ace AI | Plug and Play Batch 17

[Sam Hajj @jamx]

@bitcoinbasha @vkhosla He’s a founder. Great role model to have imo

[صندوق الاستثمارات العامة @PIFSaudi]

يرسم #صندوق_الاستثمارات_العامة مستقبلًا تقوده الاستثمارات الريادية المحفزة للابتكار، والداعمة لنمو القطاعات، والممهّدة لفرص غير محدودة.
هذا هو #تأثير_الصندوق #نستثمر_للأفضل

[Sam Hajj @jamx]

@TurnerNovak @anothercohen Literally true

[Sam Hajj @jamx]

@yacineMTB Great innovations come from giving techies power and control

[Sam Hajj @jamx]

@xpasky 😆

[Sam Hajj @jamx]

@nsbradford @ellipsis_dev @HunterFromNYC Just wanted to say that I know how hard you worked. I remember talking at 3AM on weekends.

I respect the hell out of you sir. I know you’ll go on to even greater things.

Ellipsis couldn’t have had a better CTO or better founders at its inception.

My sincerest congratulations

[Sam Hajj @jamx]

@MartinGTobias Tell her to do her own thing and be supportive

[Sam Hajj @jamx]

claude 4 sonnet is legit. As smart as gem 2.5 pro if not more, and a better agent

[Sam Hajj @jamx]

@jefielding Take them out for something fun

[Sam Hajj @jamx]

China is forcing us to create new economic and governance models: If SOTA models like O1 and Gem 2.5 Pro can be cloned within weeks, what are the new economic models that should govern AI models?

Will our claim on intelligence stay valid in the future?

[Sam Hajj @jamx]

The world witnessed a very quiet Deepseek moment a few hours ago.

Today, Google killed off Gemini 2.5 Pro Exp. This was their free tier for a model that was the TRUE state of the art (SOTA). It came out in March. This model blew everything out of the water in terms of code.

[Sam Hajj @jamx]

The same thing happened when OpenAI found out that Deepseek R1 was trained on O1's reasoning.

Empirically, the data is starting to point towards no.

[Sam Hajj @jamx]

Will our claim on intelligence stay valid in the future?

China is forcing us to think of new economic questions: If SOTA models like O1 and Gem 2.5 Pro can be cloned within weeks, what are the new economic models that should govern AI models?

[Sam Hajj @jamx]

The same thing happened when OpenAI found out that Deepseek R1 was trained on O1's reasoning.

Empirically, the data is starting to point towards no.

[Sam Hajj @jamx]

@amasad @NateAFischer Safety is an illusion