Keep on to blur preview images; turn off to show them clearly
I’ve had very little success fine tuning over gpt-oss models and very much success with qwen3 models (even the instruction versions). Not sure if this is a case of skill issue or what, but they are not as friendly to tuning
Code & LLMs. Working on document extraction: https://t.co/WREWHG95hE
I made $66,040 in October 2025. 🧑💻 CodeFast — $20.7K ⚡️ ShipFast — $16.8K 📈 DataFast — $16.3K ⭐️ TrustMRR — $8.6K 🐥 Twitter — $1.5K 🍜 Indie Page — $850 🧬 BioAge — $637 💨 Zenvoice — $236 🎞️ YouTube — $118 🌱 HabitsGarden — $109 🛡️ ByeDispute — $48 📚 WorkbookPDF — $57 💩 PoopUp — $85 + My SaaS is about to overtake my boilerplate + I added a new income stream to my little portfolio with TrustMRR
Don't trust me? Check this out 😏 https://t.co/M96I1tBkBp
For anyone getting started with deep learning in 2025 and looking to do large scale training: use JAX + Keras 3. Unless you like suffering.
Co-founder @ndea. Co-founder @arcprize. Creator of Keras and ARC-AGI. Author of 'Deep Learning with Python'.
This whole “OpenAI or Sam Altman somehow killed the whistleblower” theory strikes me as wildly improbable and silly. Is it just me, or are there literally dozens of more probable explanations for each piece of evidence that is offered by skeptics as proof that it was murder? For instance, there was blood in two rooms. Not every bullet wound is instantly fatal, and people stagger around in pain. I doubt these rooms are very large or that far away from each other. Or that he ordered DoorDash beforehand. Suicide is often a sudden, spur of the moment decision. Read about the tragic death of the brilliant AI researcher Sam Roweis to see an example of a split-second, impulsive decision. Or that the security camera wire was cut. Maybe he wanted to spare his family from seeing him kill himself on video? So they wouldn’t be traumatized by the images forever? Or maybe he was experiencing a psychotic break due to underlying psychological issues, like emerging schizophrenia or bipolar disorder, and this is what caused him to imagine wrongdoing in the first place and lodge whistleblower complaints… And if you go down that line of reasoning, then maybe because he was frustrated that people weren’t taking his delusions seriously enough, he decided to make it look suspicious to force people to investigate them. Or maybe he didn’t want his parents to blame themselves for his suicide so tried to make it look like it could be a murder. Who knows? Seeing Elon speculate on the case spurred me to write this post because I think people are not thinking logically here. I didn’t know the guy and haven’t read too many details on the case, but I’m just applying a Bayesian lens here and Occam’s razor. I did have a brilliant high school friend who later had a psychotic break and committed suicide, though, so I’ve seen the erratic and bizarre thought processes first hand. This person went from the most rational, sharp person you can imagine to someone spouting paranoid fantasies a couple years later. Also, you need to ask yourself with anything like this, what is the prior probability? Why would Sam Altman risk going to prison for life and destroying everything he has built to kill a whistleblower? And has he ever done anything remotely violent or even threatening to anyone ever? I know people like to criticize him, but he really strikes me as someone who is gentle and considerate, even if he is Machiavellian in his business dealings and exhibits an overweening ambition that people find offensive. It’s just so absurd; you would need just an unbelievably high bar of evidence to overcome such an unlikely prior. Like a confession from a paid hit man and proof that he was paid by Altman, something of that nature. But in this case, we have a bunch of facts which are explained much better by way more plausible theories. And for what it’s worth, I think he handled Tucker’s line of questioning pretty well, with empathy and respect for the grieving family despite their accusations about him. I would have been a lot more brutal about it.
Former Quant Investor, now building @lumera (formerly called Pastel Network) | My Open Source Projects: https://t.co/9qbOCDlaqM
App to do it is linked below. I love not having to share long prompts and model specifications anymore 😅 https://t.co/3KrF4JVT76
Partner @a16z AI 🤖 and twin to @omooretweets | Investor in @elevenlabsio, @krea_ai, @bfl_ml, @hedra_labs, @WaveFormsAI, @ViggleAI, & more
RT @ekuyda: AI is ready for its GUI moment. it’s crazy that we have godlike AI and people still have to share text prompts. take this Ghos…
AI Apps investing @ A16Z; A1111; Boards of Krea, Deel, Clutch, Titan, Arc Boats, Untitled, Happy Robot + more; If you’re not at the table, you’re on the menu