81–90 of 148 posts
Earlier quoted context omitted.
A few generations ago many of our brightest minds were building doomsday weapons to incinerate all life on Earth, so maybe this is progress.
Is it? Would you rather be dead or a zombie?
"Better to Die on Your Feet Than Live on Your Knees" as they say.
Earlier quoted context omitted.
Yeah, this isn't really that different than quant hedge funds hiring theoretical physics PhDs. It's not like you're actually going to be doing Yang Mills field equations to trade some stocks. They just need a guy to run some linear regressions, and can afford to find someone really really overqualified so they don't screw up.
Not sure if these people are overqualified as much as the fact that people who check the box on paper are generally unqualified. Generally good people can only really do math up to a few years below the highest level that they learned in school. If you actually need someone to do basic linear algebra you probably want somehow who took classes that went well beyond the freshman/sophomore level that everyone has.
"Our brightest minds are making people click ads" ...
Which is a good thing, since the sociology we had before clickable ads was unequivocally less scientific and useful.
Earlier quoted context omitted.
This analogy is cringe because it's overstretched. Argument: For example, are books considered pharmaceutical products (typically not quick--that would be like "delayed release" medicine) just because you get pleasure from reading them?
Not that cringe to me. Tiktok and the like are highly addictive for some people and the dopamine kicks are real and unhealthy. Reading books does not give the same release of dopamine as TikTok/Instagram.. hence why most people don’t read for 2 hours a day but many people easily spend that much time on insta/tik. Modulating behavior can cause physical withdrawals.. you don’t need a pill. I.e lovers, gambling addicts,…
Earlier quoted context omitted.
That will be cool if their really looking to expand into something useful like biochemistry r&d, similar to Google with deepmind.
TikTok brain implant?
My money is on bio-engineering “experience” into human memory.
It’s the least materially wasteful as the materials are in the process, not outputs. The outputs are not stuff but modified organic matter state.
Tiktok/bytedance already have large ML infrastructure in place (GPU clusters), and since quantum chemistry simulation is moving towards ML based simulation, they have an edge there. Open source ML simulation software such as DeePMD-kit is already seeing contributions from Tiktok employees.
In case HN readers doesn’t know, ML based MD scales O(N) and traditional DFT based simulation O(N^3) where N is number of atoms. From first hand experience: what used take weeks with traditional DFT MD on CPUs are only taking few minutes with GPUs right now if you have trained a good ML potential. They achieved near DFT accuracy.
Earlier quoted context omitted.
TikTok brain implant?
I don’t think manufacturing of computing parts (in this case, a brain implant) is the future. My money is on bio-engineering “experience” into human memory. It’s the least materially wasteful as the materials are in the process, not outputs. The outputs are not stuff but modified organic matter state.
Earlier quoted context omitted.
Yeah, this isn't really that different than quant hedge funds hiring theoretical physics PhDs. It's not like you're actually going to be doing Yang Mills field equations to trade some stocks. They just need a guy to run some linear regressions, and can afford to find someone really really overqualified so they don't screw up.
Not sure if these people are overqualified as much as the fact that people who check the box on paper are generally unqualified. Generally good people can only really do math up to a few years below the highest level that they learned in school. If you actually need someone to do basic linear algebra you probably want somehow who took classes that went well beyond the freshman/sophomore level that everyone has.
I don't see why it's necessary to have a potentially unrelated PHD to do basic linear algebra.