Dr. Griffin & Vincent Racaniello discuss outbreak of Hantavirus on a cruise ship and its potential for person-person transmission, how cessation of funding for global health by Trump administration has compromised American health and well-being & more. 📺 https://t.co/brBj09N0oZ
@nataliealund This is wrong, I am in contact with Jonathan’s vet, Joe Hollins who does indeed confirm that he is alive. The account you are citing is not Joe (who is not on X). Pm me and I can verify.
AI generated code is not an abstraction like compiler.
Compiler outputs are reproducible, deterministic and some are mathematically verified.
LLM codegen is a non-deterministic, probabilistic code generator with a human in the loop. The human must understand what it spits out.
@thorstenball Because now you own it and are responsible for it. The point of using someone else’s code or library is that they were responsible for solving that problem and maintaining that code base!
Unpopular opinion:
Unless you’re just prototyping, you should aim to understand as close to 100% of production code generated by LLMs.
Yes, all of it.
Effective mental models are still important for humans to sustainably maintain and evolve a codebase via prompting alone.
@mitsuhiko I guess what did you expect, though? Twitter awash with claims of incredible productivity gains with these tools and that code quality doesn’t matter. It feels like this is the inevitable outcome.
@thorstenball@nilskch How do you know it wasn’t compromised and inserted something to mine bitcoin? It is still your responsibility as an engineer to know what you are deploying.
PSA to my Postgres people: use a connection pooler.
Incredible article on when and why to use PgBouncer.
Includes a great explanation of how increasing direct connections leads to more contention → degraded performance. (+ benchmarks too!)
@mattyglesias There is utility but not to the extent that the companies that provide the service or the investors claim there is. There is a vast gulf between these two points.
Most genetic reports label variants as ✅ “pathogenic” or ❌ “benign.” But what patients really want to know is: what are my chances of getting the disease? 🚨 In our new Science study, we used AI + 1.3M patient records to better answer that…
If any PhD student working in computational biology is displaced by this most recent tumult and is looking for a new lab, I have room for one (or potentially two) PhD students in my group. Feel free to reach out via email.
Huge result! Not only did a polygenic risk score identify a large number of participants with prostate cancer, many of them were low risk by conventional guidelines AND had medium/high-risk disease. That has been a hard needle to thread in other cancers.