This is really exciting. One direction I'd be especially excited about is complementary but slightly different: a modular "OpenClaw for science" layer — a reusable repo of scientific skills plus lightweight operating machinery that lets agents carry a research process across long-running loops.
By skills, I mean things like curiosity/next-check selection, hypothesis decomposition, experiment design, artifact and log inspection, honest negative-result handling, and upkeep of a running experiment ledger. The ledger lets both humans and agents quickly recover context, spot bigger-picture patterns, and avoid getting stuck in dead ends.
By operating machinery, I mean lightweight coordination: recurring wakeups, status checks, context recovery. Stuff that keeps a scientific campaign moving correctly overnight or over a weekend. For many of us in academic science, the compute environment is shared HPC clusters and Slurm/PBS/etc queues rather than Docker, so an execution layer native to that world would open this up to many more labs.
Feels very complementary to the runtime/ecosystem side of Flamebird and the end-to-end AI scientist direction in NeuriCo.
@VPrasadMDMPH For some, especially young people, it's not a utilitarian QALY calculation, it's a preference. For them, an exceedingly remote shot at a cure is worth the risk of doing harm.
@VPrasadMDMPH I mostly agree, especially that doctors should be careful not to misinform the public.
But as a person in my 30s with metastatic TNBC, if SOC can't cure, I don't think it's irrational to be interested in beyond-the-SOC options. We already know what doesn't work.
@BlueObserverA@Ihornistfinkle@felixchin1@VPrasadMDMPH IMO it's more useful for advanced programmers than beginners. Easier for us to spot errors and push the answers in the right direction. Beginners can go down the wrong rabbit hole.
@SpencrGreenberg@VPrasadMDMPH This might be a useful lense for thinking about biomedical papers where we know the prior should be low of anything being beneficial
@francofor38 What's your plan to fix the portage park pool schedule?
The pool is an amazing community asset, but nobody with a family can use it because of extremely limited hours.
@DJJonnyTravieso @TamarHaspel@kprather88 The review isn't trying to answer if perfectly-used respirators are effective. It's trying to answer if real-world use is effective. That's the question that matters for policy decisions.