Updated my Github profile with all the repos my PRs have been merged to. 2,202 merged PRs across 1,014 repos. From TensorFlow to Go, to Python, n8n and Hugging Face Transformers. Thank you to everyone who's engaged with my PRs!

Today’s AI takeoff stands on decades of open research and open infrastructure. The Transformer. Backpropagation. ImageNet. PyTorch. TensorFlow. JAX. CUDA. Linux. Books, the open Internet, and millions of open-source codebases on GitHub that became training data. Without all these, today’s AI labs and companies simply would not exist.

Me in 2012 instead of learning: DSA System Design AI / ML Generative AI LLMs AI Agents RAG Vector Databases Embeddings Fine-tuning Prompt Engineering MCP AI Coding Agents Python Java C C++ Go Rust JavaScript TypeScript HTML CSS React Next.js Vue Angular Node.js Express.js NestJS Spring Boot Django FastAPI REST APIs GraphQL WebSockets SQL PostgreSQL MySQL MongoDB Redis Elasticsearch Kafka RabbitMQ Docker Kubernetes Terraform AWS Azure GCP Linux Git GitHub CI/CD DevOps Microservices Serverless Cloud Cybersecurity Data Engineering Data Science MLOps Data Analytics Blockchain Web3 Smart Contracts Solidity Edge Computing Distributed Systems Observability Prometheus Grafana OpenTelemetry Supabase Firebase Vercel Cloudflare PostgreSQL + pgvector LangChain LlamaIndex PyTorch TensorFlow Hugging Face Ollama OpenAI APIs Claude APIs Gemini APIs Model Context Protocol AI Infrastructure GPU Computing CUDA

Agents are now producing bug fixes faster than people can review them. In the time it took humans to verify 70,000 security patches, @OpenAI's agents verified 500,000 across major web browsers. At that pace, the human signing off at the end likely isn't reading most of it. Every lab keeps that human anyway. Probably not to catch the errors, but because someone has to be accountable when an agent is wrong. 👇 ------------------- That 7-to-1 gap isn't a one-off. @Google moved production models from TensorFlow to JAX at 6 to 8 times engineer speed, and on @YouTube's models, work that would have taken engineer-months finished in weeks. The agent does the volume, and a human still signs the bottom line. Nobody read 500,000 patches. The volume an agent clears in a day is more than a human gets through in a month, so the person approving Google's migration or OpenAI's security work is mostly taking it on trust. Anyone running a pipeline like this can see the same math. So why keep the human? Most likely for accountability. When something ships and breaks, someone has to own it, and "the agent approved it" isn't a sentence a company can offer a regulator or a court. The human is the name on the decision when it goes wrong, which may be the real job long after the reading stopped being possible. So, when do we go fully autonomous? It sounds like a question about agent capabilities. But on this evidence they're already good enough to outproduce the humans watching them. What's holding autonomy back is that removing the human removes the accountability, and nobody has built the replacement. A human's sign-off carries weight because a person can be held responsible. An agent's approval is just an output, with usually no record of what was checked, against what standard, that anyone can stand behind later. So companies keep a human in the loop as cover, reviewing a fraction of the work, because the alternative is an agent whose approval means nothing when it counts. Most of the field is racing to make agents more capable. Far fewer are building the thing that lets an agent's work be trusted without a human babysitting it. That's part of the infra we build at Sahara AI: verifiable execution and an unforgeable record of what an agent did and what it checked, so its sign-off can be held to account instead of taken on faith. Follow @SaharaAI to keep up with agentic AI, the infrastructure it demands, and where the real building is happening. About Sahara AI: Sahara AI is the agentic AI company dedicated to making AI more accessible and equitable. We build the core protocols, infrastructure, and applications that let personal agents anticipate and execute on your behalf. For this to work, infrastructure has to be trustworthy: verifiable execution, enforceable usage policies, and automatic value distribution across every tool, model, and service an agent touches. Sahara is building a growing suite of agent-powered applications on top of this foundation, including @HeySorinAI, your personal agent for global digital markets. Our solutions currently power AI agents and high-quality data for consumers, Fortune 500 enterprises, and leading research labs, including @Microsoft, @Amazon, @MIT, Motherson, and @Snap.

Why is the Midjourney team uniquely qualified to build medical hardware? @BeffJezos, founder of @extropic: "David's a trained physicist. He worked at NASA. He was a founder of Leap Motion, which is completely slept on as one of the first major projects he worked on." "What I really like about that team is that a lot of the same team from Leap Motion followed David into Midjourney and now into this hardware exploration." "They were training neural nets before there was TensorFlow, before there was GPUs. They did it on CPUs, and they had some of the first neural network technology deployed in a consumer product." "So people are like, 'Why would a gen AI company pivot into medical devices?' It's really folks that came from the hardware world finding this unique blend between neural nets and hardware to achieve things people didn't think were possible." "If it's physically possible, I very much have faith that that team can pull it off."

TensorFold 1.2.6 is out. 🚀 The big one here is the prompt cache under multi-agent load. In the 16-agent Qwen3.8-27B + DFlash2 test, median follow-up time to first token dropped: • M3 Ultra: 546s → 15.9s • GB10: 105s → 25.8s Each agent started with a ~17k-token prompt. Same replies throughout that test. Also in this release: • Flash Next cold prompts finish 29 to 41% sooner on HIP, measured on the Radeon 8060S. • GLM-5.3-Flash single-stream code decoding is 32% faster on two GB10s at TP=2. • Opt-in DSpine-b16 drafting from Tencent for Qwen3.8-27B on CUDA. • An image-generation API for image-capable engines, local Unix sockets and tool-call fixes. One upgrade note: kept prompt states need a fresh prefill after updating. Massive thanks to everyone contributing code, testing on their hardware and reporting the things that need fixing. I really appreciate the time you all put into this. @andersjw_ @Bizuayeu @defilan @devopsfortunato @unigilby @jvr0x @teleogen And on GitHub: abm9111, AbrahamCete, akol1, BobClawblaw, chadhurley25075-png, davidtai, ElderOrb, jschmied, kotevcode, Mirrdhyn, mrv777, ppardus, robertung88 and volfco. @petruspennanen @volatilemarkts @plotarmordev @MiaAI_lab

TensorFold is coming to @ComfyUI. 🔥 I posted about wanting to integrate TF into ComfyUI, and they replied, "Let's talk". We're starting with an experimental MOSS Image sampler on GB10, using TensorFold's native CUDA worker. Video will be coming shortly. Huge thanks to the amazing Comfy team. I'm really excited to push this further. PR:

TensorFold 1.2.5 is out. 🚀 This release brings: • TensorFold Web, a local browser app for chat, models, learning, benchmarks and more. • LFM2.5-24B-A2B on CUDA. • Nemotron Parse and LightOnOCR-3 on HIP. • Clef's joint schema head on CUDA. Qwen3.8-27B MLX 8-bit on M5 Ultra also gets faster. In the measured test, cold prompts were 18% faster and decode 7.6% faster, with the same tokens and prompt state. There's lower memory use for Qwen CUDA streams, better conversation-cache handling and a fix for PDF rendering crashes on macOS too. A massive thank you to everyone contributing code, testing builds and reporting problems. These releases take a lot of work, and I really appreciate the time people put into them. @Bizuayeu @devopsfortunato @unigilby @teleogen @dubsthedev @kalin_swiss @Q4_K_M And on GitHub: AbrahamCete, gordonhatusupy, jschmied, mrv777 and yguanrong. Special credit to Stephane on GitHub for TensorFold Studio, which TensorFold Web builds on. @volatilemarkts @plotarmordev @MiaAI_lab @petruspennanen

TensorFold 1.2.4 is out. 🚀 Two of the measured wins in this release: • DeepSeek-V4.1 restored a saved 319K-token conversation from NVMe in 1.2 to 1.4 seconds, versus 155 to 178 seconds, on two GB10s at CUDA TP=2 with replay prefill. • Flash Next on M5 Ultra: two overlapping clients finished in 47 seconds instead of 127 in the MLX 4-bit parallel-serving test. Also new: Gemma 4 E4B on CUDA, LFM3-1.7B text on Metal and Clef-Flash decisions on HIP. The HIP runtime now builds and loads on Windows too. There are cold-loading improvements, tool-call fixes and clearer API errors. One upgrade note: existing disk-cached prompt states need a fresh prefill after updating. Thank you to everyone behind the code, testing and bug reports. The community puts an incredible amount of work into these releases. @arminxnova @atasrb @Bizuayeu @BlivionIaG @codengod @defilan @devopsfortunato @unigilby @jayleaton @jishvenugopal @Wi11y010 @teleogen @lasse_ai @petruspennanen @dubsthedev @woolsey And on GitHub: AbrahamCete, AdrianBinDC, BobClawblaw, ElderOrb, gordonhatusupy, goshi, jschmied, laidick, ppardus, shuuul and tolewis. @petruspennanen @plotarmordev @volatilemarkts @MiaAI_lab

TensorFold 1.2.3 is out. 🚀 Some useful Flash Next improvements in this one: • NVFP4 on two GB10s at TP=2: plain decode 21% to 23% faster, with 2.62 GiB less model memory per rank. • M3 Ultra: a 1,024-token reply beside a 64K prompt finished in 8 seconds instead of 48 in the parallel-serving test. Same tokens. Those are specific test results, with the configurations and comparisons in the release notes. We've also added PDF document input, Ornith-1.5-122B on Metal, native dots.mocr OCR and Qwen embeddings/reranking on HIP, and LFM3-1.7B on Intel Iris Xe. There are fixes for prompt caching, memory use, streamed tool calls and serving too. A huge thank you to everyone contributing code, testing builds and reporting problems. There's so much work behind each release. @arminxnova @Bizuayeu @BlivionIaG @codengod @devopsfortunato @diwakergupta @unigilby @jayleaton @jishvenugopal @JRB_AI @joshuaswarren @teleogen @lasse_ai @MiaAI_lab @woolsey @petruspennanen And on GitHub: AbrahamCete, BadAd84, BobClawblaw, dinadur, garvrao80, joernchen, jschmied, ppardus, shuuul and tolewis. @MiaAI_lab @volatilemarkts @petruspennanen @plotarmordev