chermi

31 ms·

451.
by chermi 1y ago
I suggest you read about the history of asml, where the tech came from...
452.
by chermi 1y ago
Ah yes, ban and regulate your way to innovation. That's certainly going to work. And of course, you threw in Reagan. You'll certainly get upvotes, but nothing else.
453.
by chermi 1y ago
Ok I guess we're just redefining things now. Btw, NYT is now far left. See how easy that is?
454.
by chermi 1y ago
I don't understand how he's at fault for the field being behind where it maybe could've been, especially the language "held back"? Did he actively discourage people in against trying his ideas as compute grew?
455.
by chermi 1y ago
How cool, it's still in the ACS logo! Germans have the best names lol. Calibration apparatus.
456.
by chermi 1y ago
I prompt it to come up with 3 potential implementation plans, choose which one it thinks is best, and explain its plan to me before it does anything. I also ask it to enumerate which files/functions it will modify in its chosen plan. T
457.
by chermi 1y ago
You're right. I was focused on the idea that for software to be taxed as a capital investment there had to be a time when it was considered a finished product. Like building a tractor. I guess the analogy then is how the tractor builde
458.
by chermi 1y ago
Thus firmly placing this in the regulatory capture category.
459.
by chermi 1y ago
Thanks, this is the closest thing to making sense. But it still doesn't make sense. Like you said, this is a pretty weird characterization of software. I guess it would make sense to lawmakers who have no idea how it works. Combine th
460.
by chermi 1y ago
While I largely agree, I don't think it's quite correct to say AI generated blogs contain no new information. At least not in a practical sense. The output is a function of the the LLM and the prompt. The output contains new infor
461.
by chermi 1y ago
Can someone steelman the positives for me? I don't see how it's anything but pure regulatory capture favoring established tech firms. A small company ramping up revenue simply can't handle this amortization while a large, est
462.
by chermi 1y ago
Maybe I should just read the wiki you linked, but I guess I'm confused on how this is different than steepest descent? I'm a physicist by training so maybe we just call it something different?
463.
by chermi 1y ago
Kind of, but not really. I would instead say that the idea of "stickiness" or trying to make your software sticky is rent-seeking. Of course, most SaaS companies vigorously pursue stickiness, but it is not an intrinsic property of
464.
by chermi 1y ago
Ahh I see. Yeah, we really screwed the pooch on standard nuclear. We knowingly forced ourselves off the learning curve, and now people point at that as evidence it doesn't exist for nuclear.
465.
by chermi 1y ago
The first of anything will always be expensive. The real test is whether we get good learning curves for SMRs, that's (to me) their main advantage.
466.
by chermi 1y ago
I know. But you should realize that sometimes opinions can be wrong.
467.
by chermi 1y ago
Lol the authority you speak with is so hard to reconcile with the absurdity of your claims.
468.
by chermi 1y ago
I thought it helped with cravings/addictive behavior. Which is the opposite of you're saying, if we're all accepting in this post that processed foods are designed to work on cravings and are addictive.
469.
by chermi 1y ago
"...a note should capture just one thing, and if possible, all of that thing.". This just doesn't seem feasible to me for an scientific workflow. I realize his scope is math, where I can picture it being a little easier, but
470.
by chermi 1y ago
Good point. I'd be curious if anyone actually has the text showing he said this. It's in principia I guess. My bet was that he never gave a full description, but rather just said that it is moon/sun that *causes* the tides.-
471.
by chermi 1y ago
Awesome, thanks! It's interesting how the most effective LLM use for coding kind of enforces good design principles. It feels like good architects/designers are going to be more important than ever. Edit- Except maybe TDD? Which
472.
by chermi 1y ago
TL;DR newton basically got the FORCES right, but forces don't tell the whole story because of (mainly ) 1) insufficient propagation speed because ocean is deep 2) think of it kind of like a diff eq, the boundary conditions (largely fr
473.
by chermi 1y ago
I'd interested to hear more about your workflow. I use Gemini for discussing the codebase, making ADR entries based on discussion, ticket generation, documenting the code like module descriptions and use cases+examples, and coming up
474.
by chermi 1y ago
My understanding is that the ability to watch the chain of thought is no walled behind the ultra subscription?
475.
by chermi 1y ago
Protein structure prediction is a pretty useful tool. There are various 'foundation' models in biology now that are quite useful. I don't know if you want to count those as AI or ML. If you're looking for breakthroughs d
476.
by chermi 1y ago
I think maybe they are talking about the very tightly packed yet still functionally accessible 3d structure that is chromatin, not individual strands.
477.
by chermi 1y ago
I see, apologies if I was too aggressive in my reply.
478.
by chermi 1y ago
Oh sure, I'm all for that in the mean time. But the people funding this are looking for big payoff. I want to be clear that this is not my field and I'm probably a bit behind on the latest, especially on the algorithmic side. IIRC
479.
by chermi 1y ago
Again, I'd say based on the conditions for the theory to hold, the US couldn't prove or disprove the theory either way. I guess to summarize my stance: we don't know what the learning curve for nuclear is because data that is
480.
by chermi 1y ago
Based on the comments in this thread... Guys, Microsoft fuckery doesn't invalidate an entire field. I think certain VCs are a little too optimistic about quantum computing timelines, but that doesn't mean it's not steadily pr

More ›