We’ve just made public 125,091 DICOM radiology images from across my cancer journey (link below). The data is both longitudinal and multi-modal, dated and linked to my treatment timeline, so you can compare imaging across initial diagnosis, standard of care, first remission, “founder mode cancer care”, and second remission. It’s viewable in OHIF, the standard open-source viewer used in radiology research. Given the long history of applying deep learning to radiology, we’re hoping to see a variety of methods and models applied to my data. If you find something interesting, please reach out to cancer@sytse.com, or if you know a researcher who should see this, tag them below. Thanks to Terra LaFerriere for organizing these images, and Kamil Slowikowski for building osteosarc.com into the resource it is.

#BREAKING Special Executive Magistrate Court extends CJP leaders Saurav Das and Ashutosh Ranka’s judicial custody for two more days. @Cockroachisback @SauravDassss @AshutoshRanka

I write briefly about my ongoing recovery from a viral infection. nihalsahu.net/ageing/

Coin labs is announcing our new decision model Flip-1. - Distilled down to 1 parameter - 50% on yes/no bench - Runs fully offline

Every few years there is a new current thing in tech. Sometimes, India will fall behind, and whenever that happens, everyone around: those in the government or adjacent (ministers, regulators, other bureaucrats, etc.), those in the industry (investors, founders, engineers, etc.), and miscellaneous thought leaders (billionaires, policy think-tank people, influencers) all inevitably panic and start asking how India will catch up. Today, the current thing is AI, India has fallen behind and the same questions are coming up again. With this begins a very predictable sequence of events. The whole ecosystem goes through the seven stages of tech grief. 1. Disbelief ("India is obviously the IT capital of the world"), 2. Denial ("India has cheap labour, so we don't need AI"), 3. Anger ("All because of bloody brain-drain and Indian founders just doing D2C and grocery delivery"), 4. Bargaining ("India should build on top of existing AI, be the use-case capital of the world"), 5. Depression ("Maybe we should all just move to SF, invest in founders with a US visa or in American companies"), 6. Testing (India AI Summit, India AI Mission, Government starts investing in companies like Sarvam, funds invest in international brown people), 7. Acceptance ("Indic AI will be the global frontier tech leader of the Global South"). The saga ends with everyone coping, and hoping India will be a different place soon, fully ready for the next wave. But nothing much changes. The real answer, especially for a democracy with India's scale, is that if you know you need to catch up, it's already too late. Meta can go founder-mode, spend big, hire big and make a great attempt. China can redirect a bunch of human capital (smart people) and corporate capital (balance sheets) to solve it and get close too. But what do you do in a country like India? What the India should do is let the market, with all its messy constituents: individuals, founders, talent, institutions, universities, investors; both good and bad -- run free. The most well-meaning panel of bureaucrats cannot optimally allocate compute, a conference of regulators and incumbents will not bring frontier talent to the country, government handouts to already quite large funds are a waste of taxpayer money, and governments investing in early stage companies causes premature king-making and throttles competition. The State can do a lot, but it has start with doing less. What does that mean? It means taking a step back, and removing all hurdles to increase the flow of capital, talent, and risk. It's the only way India will retain capital, technical talent, and risk-takers for long enough for compounding to kick in, and benefit from it as a country. You make capitalism easier. This includes reducing the compliance burden pushed onto a company (further outsourcing to CAs and lawyers doesn't count), but also making it easier to raise as a company, easier to invest as an investor, reducing paperwork for everyone (digital disclosure forms are also paperwork), etc. Startups are not unique kinds of entities requiring any special laws, registration and benefits. Startups are just companies. The same kind of ease of doing business should apply for everyone. If you're worried about tax evasion, remove all tax breaks. When it matters the most, startups don't make a profit anyway. Will some bad apples try to game the system and get away with it? Sure. There will always be bad apples. You let it go, because the long-term benefits outweigh all short term costs by orders of magnitude. Even the IPL -- our one crown jewel in the world of sport, had corruption, match fixing, auction rigging, and a bunch of other shady business early on. You fix it later. No sane person would go back in time and shut down IPL before the second season because of problems at the time. Innovation is delicate. If we think in decades, progress will follow. Many of these things are already solved elsewhere, there are global benchmarks we should strive to beat. Our competition is not just an India of 20 years ago, but every other country we lose innovation to. We are losing, make it easier for those who will help us win. In case this is too vague, here is what we should aim for: 1) bring down company incorporation time to <1 week, 2) bring down time to close investments <1 day, 3) bring down time to open a corporate bank accounts to <1 day, 4) minimise the total number of hours a month a promoter has to spend in non-tax compliance, 5) bring down the number of hurdles to export goods and services. Part 2 to follow soon, where I will try to list out a hundred small things that will help fix this. My DMs & comments below are open: for any recommendations from those facing adjacent problems, for anyone who would want to collaborate on these, and especially for anyone inside the Government interested in our feedback.

P.S. AI protein design gets the headlines. Whether the thing in the tube is actually what you designed gets decided by analytical methods. My team at @PopVaxIndia is hiring an analytical scientist for protein characterisation (SEC-HPLC) jobs.popvax.com/41532

What a bizarre future. I wanted to play Halo: Combat Evolved for the first time with @robertskmiles, and we wanted local multiplayer across two screens, and a model could just immediately add that feature to the decompilation.