Adam Louly

@AnassAb01 @BankAlMaghrib

Hello @BankAlMaghrib, the team at Optiware is ready to build it and maintain it for the next 2 years for free. We accept any conditions and security requirements you have. Just make this possible I need it, my friends need it and the whole fintech industry needs it.

@Ghitz01 99% dial phds makhssh ikoun m3tarf bihom

Saraha PhD dial 3 ans makhssh ikoun m3tarf bih

@arbaoui_mehdi Had i7sas kan kit3ach f 2009-2010 machi f 2026

ماشي نفس الأورڭازم ديال تكتب كود راسك ولا تخلي كلود يكتبو ليك. يقدرو يوصلك لنفس النتيجة، ولكن معمرهم يوصلك لنفس الإحساس.

@isbkch Don’t have a number yet, probably 0 if we’re talking about money

@LoulyAdam Depends... What's the ROI on this?

AI bill for this week ... this is getting out of hand.

@alilakrakbi Inconvenient even f states, takes 40 mins for fast charge to 80%. Maybe have two: one electric and one gas that you can use if you’re short in time

كنعيا ما نقول ليكم الحل هو طوموبيل الضوء.

@isbkch M4 Max 36GB is doing bad too! Thinking of getting the 128GB one 👀

I have never thought the M4 Pro 24 Go of RAM will do me this bad Looking at you Superset AI

@zakelfassi Glisat lmla7

iykyk

@danielhanchen Wait does that mean you pay more if this is enabled? 😅

If you find Claude Code with local models to be 90% slower, it's because CC prepends some attribution headers, and this changes per message causing it to invalidate the entire prompt cache / KV cache. So generation becomes O(N^2) not O(N) for LLMs.

@real_deep_ml I love to see how AI is reshaping education! great work!

Just finished adding all the questions based on this book. Still need to create all the playgrounds and a collection for all the questions, but you can start solving them while reading this great book

How do you entertain yourself while waiting for your codex-5.4-X-Mega-High-Thinking to finish its task?

@adilmania What do you use to create this? Looks dope!

february was absolutely mogged.

@soub4i I don't think kprintf exists in cuda, I've never used it before. apparently the printf fix was a false positive and ended up being a totally different issue.

@LoulyAdam What happened to the kprintf() 👀 .. but for answer your question: No!

Is it ever legal to leave a printf in a cuda kernel if it resolves a heisenbug?

Bullish on @cerebras

@MIduoad I think he got a point, how would you cluster ga3 dok ideas and come up with top k ideas?

He lost me when said daka2 l stina3i 😂😂

@MartinShkreli @elonmusk Stick to pharma, binaries generation is much cheaper on cpu

everyone laughed when @elonmusk said something like this for writing binaries. i guess we're here!

Is there a way to have one agent act as an orchestrator so you only talk to that agent, and it then manages the other 4-5 agents, gives you status updates on each, and dispatches tasks to them?

That’s not accurate. The $500k investment from YC wouldn’t attract a group of ex-FAANG, that’s probably less than their TC.. Unless it’s for the network, I believe YC invests more in ambitious young founders, which is what I appreciate about them. Unlike A12Z, Sequoia, those firms would invest in experienced founders from major tech companies.

@LoulyAdam @ycombinator @llmgateway Yeah, it seems so. The investment in retarded people working and coming from big corporations is a way to synchronize their money flow without adding any value to the market 😅

I’m genuinely curious to understand how @ycombinator selects startups for investment. - Real users? @llmgateway already has a significant number of real users who pay real money to use their service. - Good revenue? If trustmrr.com/startup/llm-ga… isn’t sufficient, you can contact @smakosh, and he will share a screen with you to view @stripe’s revenue. - Well-educated founders? Their testimonials already demonstrate their genuine behavior and characteristics. - Project potential? It’s either the current or perhaps the only startup in the market that faces openrouter (or we can call them closedrouter 😭) competition.