This is the AI Agent I built for the @NVIDIAAIDev Contest. I got top 100! #AIAgent #ArtificialInteligence . Thank you @AndrewYNg @isafulf & @LangChainAI . I've given away all the code for free for your use, even commercially for your profit if you want.

Get yourselves a mentor who cares enough about you to hurt your feelings to turn you from what you are, into what you could be.

When you work for 4 months on a project, ship, succeed and you’re waiting to see Day 1 of MRR begin! …and August 1st shows up but it’s not there 😭 Client cancelled 😞 before it came. Andreesen was right about Founders experiencing only two emotions.

RT @kimmonismus: This is so freaking amazing, holy moly. Now give it a Voice Mode and we are good to go.

Hasn’t everybody? Show me the over 100 club and I’m impressed. 80-90 should get the job done though.

4o changed under the hood without OpenAI telling us again.

36, you guys?.

The umbrella I was holding fell at the end there, but it still hasn’t hit all the islands at this point.

Do you remember when you joined X? I do! #MyXAnniversary. Wish we weren’t celebrating with a #Tsunami in Hawaii though…

With the 8.7 Earthquake off Russian coast, Hawaii expects 9 foot higher waves 🌊 this evening! Good thing I found this box, 10 feet above the shore, to watch from! #Tsunamiwarning

I’ve been really impressed with the conversations @DavidSacks and the besties had in 2021. You mentioned that the current Apple would not have hired Steve Jobs. I ran across this in a 2011 magazine article on him. AGM had a great take I agree, but this is also true, save Atari.

Does anyone have a podcast to suggest in the business tech side similar to BG2, Founders, The Memo, Smart Friends or a16z?. I’m surprised just how much of an advantage I’m getting from these. Are my followers podcast listeners?.

Hey @OpenAI your 4o model is embedding it's CoT sequence too deeply inside its own analyzer and needs a higher level agent oversight. It hallucinates on first sigma events which means you over-agentized it on the code environment but not the top level. This is why it breaks.

Nice job @isafulf on the Agent video from OpenAI, you’ve come pretty far from when I first watched your videos with Andrew on prompt engineering.

Hello! @MistralAI I tried running your devstral-medium-2507 in place of GPT4.1. You have great speed, but I’m getting agent breaking hallucinations. What agentic benchmarks did you outperform 4.1 on again?. I have a production system operating in two business and periodically I.

I've spent 90 days testing over 40 LLM models to handle my agentic requirements. Only @OpenAI's GPT 4.1 is able to handle everything I needed without hallucinations from my SYSTEM_PROMPT. There are many AI Agent Engineers out there, I hope this 3 months of discovery helps. Those.

Have there been any developments @friedberg on your 0.1% stock trade idea which equal the 165 billion in capital gains you talked about on the GME Podcast in 2021?.

RT @naval: If AI lets non-developers replace junior developers, imagine what it lets junior developers do.

I couldn’t be more impressed with Brad Gerstner from @altcap and his movement Invest America! 🇺🇸 This is going to do more for our country than almost anything else over the next 50 years+! . Thanks @bgurley and of course @MichaelDell for the latest BG2 Podcast!.

Is anyone else working on PCI DSS Level 1 payment processing through agentic systems?. I mean, I know VISA is starting on it, but anyone else in my circle doing this? My system is almost done, but with 1 glaring flaw and I don’t even think Sierra is doing this yet.