Harsh Gupta

pov: you’re asking if india is building the next chatgpt. short answer? probably not. and that’s okay. because the real value isn’t there. while silicon valley chases bigger models and parameters, indian builders are solving problems that actually need solving. a doctor managing 900 patients. a farmer waiting for advice in their own language. a founder dealing with regulations that don’t copy-paste from the west. this work isn’t glamorous. it’s vertical ai. and when constraints are part of the system, this is the only path that makes sense. look at the math. healthcare gaps keep growing. agriculture supports 42% of the workforce, with rising losses. vertical ai can unlock 25–50% of employee work value. generic platforms barely touch 1–5%. and it happens fast. weeks, not quarters. last year says a lot. 17 indic and vertical models. genai assistants grew 4x. services startups scaled 4.6x. yet we keep asking the wrong question. “can india build the next chatgpt?” the better one is simpler. can we build a few vertical ai companies that solve real problems and people are happy to pay for? you don’t need to own the foundation model. own the bridge between what models can do and what customers need. that’s the real advantage. and that’s where things get interesting…

28

26 Comments

Soumith Chintala, 𝗰𝗿𝗲𝗮𝘁𝗼𝗿 of PyTorch and one of the most respected builders in 𝗺𝗼𝗱𝗲𝗿𝗻 𝗔𝗜, has joined Mira Murati ’s fast-rising startup, Thinking Machines Lab. After 11 years at Meta (Facebook), where PyTorch grew from a 𝗻𝗶𝗰𝗵𝗲 𝗽𝗿𝗼𝗷𝗲𝗰𝘁 𝘁𝗼 𝗽𝗼𝘄𝗲𝗿𝗶𝗻𝗴 𝗴𝗹𝗼𝗯𝗮𝗹 𝗔𝗜 𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻, Chintala’s move signals a new wave in human–𝗔𝗜 𝗰𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝗼𝗻. Thinking Machines Lab isn’t just gathering top talent from Meta, OpenAI, Anthropic, and leading universities—they’re offering 𝗴𝗮𝗺𝗲-𝗰𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝘀𝗮𝗹𝗮𝗿𝗶𝗲𝘀 (up to $500K), 𝗮 𝗺𝗮𝘀𝘀𝗶𝘃𝗲 $𝟮 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 𝘀𝗲𝗲𝗱 𝗿𝗼𝘂𝗻𝗱, and are aiming for a $50 billion market valuation. Their new product, Tinker, is already in use at Princeton, Stanford, and by major enterprises. Meta, meanwhile, is regrouping under the 𝗻𝗲𝘄 𝗦𝘂𝗽𝗲𝗿𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗟𝗮𝗯𝘀 division led by Alexandr Wang, with other AI leaders reportedly on the move. Chintala’s reasoning: time to pursue “𝙨𝙤𝙢𝙚𝙩𝙝𝙞𝙣𝙜 𝙨𝙢𝙖𝙡𝙡, 𝙨𝙤𝙢𝙚𝙩𝙝𝙞𝙣𝙜 𝙣𝙚𝙬, 𝙨𝙤𝙢𝙚𝙩𝙝𝙞𝙣𝙜 𝙪𝙣𝙘𝙤𝙢𝙛𝙤𝙧𝙩𝙖𝙗𝙡𝙚.” His legacy—the PyTorch tool now thriving in classrooms from 𝗠𝗜𝗧 𝘁𝗼 𝗿𝘂𝗿𝗮𝗹 𝗜𝗻𝗱𝗶𝗮—𝗶𝘀 𝘀𝗲𝗰𝘂𝗿𝗲. #AI #TechNews #PyTorch #Meta #ThinkingMachines #Startup #AIJobs #TalentWar #Innovation #BigFunding

8

I love seeing ML teams who refuse to treat biology as an afterthought. Relation ‘s PatchDNA work is one of the clearest examples this year of what happens when you start with how biology actually works and only then design the model. No shortcuts, no flattening of the problem just to fit the tooling. That mindset hits close to home for me. At Outpost Bio, we’re obsessed with the same question: how do you build AI that doesn’t attempt to overwrite biology, but explains it? When you’re dealing with trillions of microbes shaping what we absorb, how drugs behave, and how our bodies respond to the world (biology’s “hidden majority” that most research ignores), you can’t get away with convenient abstractions. You have to let the system speak for itself. PatchDNA is a great example of doing that well: a biologically grounded inductive bias for DNA that gives the model structure instead of asking it to guess. Huge congrats to the Relation + Rosalind teams on the NeurIPS Best Paper. This is the direction the field needs: ML that respects mechanism. Congratulations! Cristian Regep Alice Del Vecchio Lindsay Edwards Rosie Rodriguez PhD Jake P. Taylor-King Sara-Jane Dunn #NeurIPS2025 #NeurIPS 🏆 Best Paper Award: NeurIPS AI4D3 Workshop PatchDNA: A flexible and biologically informed alternative to tokenization for DNA https://lnkd.in/dPHswTGP 🎤 Talk: Saturday, December 6th, 11:30 Convention Center, Upper Level, Room 28A–E 🖼️ Poster Session: Sunday, December 7th, 15:55–17:00 AI4Science Workshop, Convention Center, Upper Level, Ballroom 20A

85

4 Comments

What we often chalk up to “expected randomness” in LLMs turns out to be partly a byproduct of infrastructure choices—batch sizes, request load, how attention kernels are implemented. The idea of batch invariance stood out. If your kernel changes reduction order when batch size or KV cache state changes, then your “same prompt” no longer means “same output.” For applications that depend on reproducibility, audit trails, or consistency (e.g. RL training, regulated domains), that’s not just academic. What do you think: for your stack, how big is the gap between “probabilistic slack acceptable” vs “exact repeatability required”? Are we shifting toward expecting determinism even in deployed systems? https://lnkd.in/gbKTt9-g #AIEngineering #Reproducibility #LLMInference

12

Introducing Agentic Culture — an open-source experiment in collective memory and in-context cultural accumulation for multi-agent systems. A hot topic from the recent Andrej karpathy // dwarkesh podcast is that LLMs don't yet have culture. >>> So we built the scaffolding for them to develop one. Agno Agents can now tap into a shared cultural memory — knowledge that persists beyond individual sessions, users or memories. Culture is collective, evolving context that shapes agent behavior over time. Now they can create, read, explore and learn from their culture. You can use the `CultureManager` class to create and manage `CulturalKnowledge` objects. These are stored in your database of choice (SQLite for now) and automatically retrieved as in-context cultural grounding for your agents. To give your agents access to shared culture, just set `add_culture_to_context=True`. That's it. Agents now learn from collective experience. Explore the code and README: https://lnkd.in/e7E3AcJX Agno is open source — come build with us: https://agno.link/gh

98

4 Comments

The recommendations of the Why LLMs Hallucinate reminded me of Prof SAK’s course on Formal Methods of Program Verification when I was an undergrad at IIT Delhi. The content of the course was great fun to extremely mathematical people but the grading was hell. His surprise quizzes awarded +1 for the correct answer and -2 for the wrong answer. The people in the class mostly had top 150 ranks in the JEE, and very high GPAs. Most had spent their last few years obtaining extremely certain answers with perfect reasoning. These were not people you would accuse of confident guessing - the JEE used to be dominated long form symbolic and numerical problems. The highest score in one of the quizzes in this course was +2 out of 20. I write “+” because the average score was -2, minus two. Some students missed the quiz. They had the above average score of zero… Thirty years later, this course still features in our conversations. Excessive penalties on confident guessing make for great stories but rewarding abstention with above average scores is not a good idea. I suspect we will end up hobbling models into analysis paralysis.

64

4 Comments