Personal blog of Ankit Gupta
“AI for bio” is getting hot again. Given the excitement in the current moment, I thought I’d share a bit about what actually makes biology uniquely hard as an application domain for machine learning. The reason is not simply that biology is complicated, though it obviously is. ML is good at many things that are complicated. The deeper reason is that drug discovery does not have the kind of clean…
Early users are a strange bunch. Here is a worthwhile thought experiment: how many products do you use that you were one of the first 10 users of? I would bet that for nearly everyone that number is 0. Most people are not early adopters. There is nearly nothing any startup founder could say that could lead them to be their first customer, and especially not a paying customer. There are other…
Most biotech startups are default dead. In fact, most of them earn zero revenue throughout their entire existence, regardless of how many millions of dollars they raise. This is usually true even after they IPO – most are still many years from being able to sell their therapeutics on the market. For software companies, being default alive vs default dead has a simple definition as articulated by…
Microsoft AI released a report this week about MAI-DxO , their research on an LLM orchestration system for medical diagnosis, and an accompanying blog titled The Path to Medical Superintelligence . To me, this is one of the most exciting application domains of AI. Here are some thoughts on Microsoft’s approach. These are a mix of my own opinions and reactions from some friends who are doctors. The…
These days, I’ve had the privilege of a courtside seat to the radical transformation happening in our society due to artificial intelligence. It has been awe-inspiring and I’m writing this mostly because outside of the Silicon Valley tech bubble it seems as though no one in a position to prepare our society for this is remotely aware of what’s coming. Worse, the trusted arbiters of information in…
I’m reviving my blog after some time away — it’s been an eventful 12 months. In February 2024, Reverie Labs, the startup I co-founded in 2017, was acquired by Ginkgo Bioworks. I’m now on leave from Ginkgo and I’ve joined Y Combinator as a Visiting Partner, giving me the chance to work with the next generation of companies. Especially in this new role, I’ve been thinking a bit about what worked,…
I’ve recently been spending some time getting large language models to work on my no-longer-excellent TITAN V GPU. This is the consumer-grade version of the V100 (it’s basically a lower memory version of it), running on a Ubuntu box with a reasonably powerful 20-core Intel Core i9 CPU from a few years ago. The folks at PyTorch recently released a package called GPT-Fast , enabling accelerated…
Welcome back to my blog! I haven’t actually posted anything in several years, in part because I’ve been able to scratch this itch via my company blog whenever I’ve felt the desire to share things. That said, I’ve felt inspired recently by the amount that I see industry leaders posting about their explorations on personal technical blogs, and so I figured I would get this going again. In any case,…
I have an exciting development: I decided to join some friends in starting a company! We are Reverie Labs , and we are using machine learning to accelerate drug discovery. My co-founders Jonah and Connor have been working on this idea for a few months, and I have now joined them full-time. We are using a variety of machine learning techniques to build models of chemical and biological phenomena,…
Two recent publications, both in review at ICLR 2017, propose methods to do unsupervised machine translation. These works propose overlapping methods to translate between languages without any parallel corpora . To me, at least on first glance, this result seems highly unintuitive. So, I’m going to use this blog post to summarize these two approaches, and discuss their high-level differences. Note…