bigstrat2003

一昨日議論

There are plenty of shops where they are requiring people to use LLMs. Not "we require you to produce work at X rate" such that one can't hit the target without an LLM, but actually mandating use of LLMs based on the (unproven) assumption that it will boost productivity.

bigstrat2003

一昨日議論

Opportunity cost is not a real cost. No money leaves your bank account. It is a decision making tool, and treating it as an actual financial cost is a misuse of the idea.

bigstrat2003

一昨日議論

Yes, pretty much.

bigstrat2003

3 日前議論

The problem with RHEL is that the free version exists at the sufferance of IBM, a company not well known for being motivated by keeping tech enthusiasts happy. I would use Debian, personally. Not as long of a support cycle as RHEL, but still quite stable and no possibility of corporate rug-pulls.

bigstrat2003

3 日前議論

I would kill (figuratively, anyway) to have no AI whatsoever. No slop machine threatening to replace my job, or turn my job into babysitting its stupidity, and hardware would be reasonably priced? That would be awesome. AI has brought me nothing but downside.

bigstrat2003

3 日前議論

> And to be fair, lots of grifters have done the fake it till you make it act on HN...

Indeed. That is how the entire AI industry exists.

bigstrat2003

3 日前議論

I once rented a VW which beeped the seat belt warning at me from the weight of my phone in the passenger seat. It was absolutely insane. Seat belt nags are an utterly terrible feature.

bigstrat2003

3 日前議論

Everyone dies eventually, so that means very little by itself. Did they die young? How much younger than their peers? Did they gain enough enjoyment from drinking and smoking to offset the fewer years of life (subjective, of course, but important to consider)?

Every intervention has a cost. Not all results will be worth the cost we pay for them. I think we are well past the point in Western society where we are giving up too much in order for too small a return, personally. With cars, but also just in general, people have lost the ability to do cost-benefit analysis and act as though every safety improvement is an unalloyed good. But they aren't, and I think that this thinking is making all of our lives significantly worse.

bigstrat2003

3 日前議論

My wife's Volvo XC40 Recharge is the first Volvo we've ever had, and it's certainly the last I'll ever have (can't speak for her). The software is so flaky. The map screen on the dash doesn't load sometimes, the side mirrors don't tilt down (which they are set to do) when reversing half the time, and every so often the sound completely stops working in the car without a hard reboot of the info system. To make matters worse, it turns out that car operation signals (like turn signal clicks) play through the stereo, so when that happens you are driving without important audio cues because they were too damn cheap to put electronic clickers in the dash.

It's probably the worst car I've ever owned, worse even than cars I got for way cheaper than the $50-60k they wanted for this thing. Never again, fuck Volvo.

bigstrat2003

3 日前議論

She's a corporate executive. Maybe she's not willing to crash the business and take a golden parachute, but if so she's one of a vanishingly small crowd. I would personally bet that she's willing to slash and burn just like other corpos are.

bigstrat2003

4 日前議論

I would still rather have that than let some game run in ring 0 TBH.

bigstrat2003

4 日前議論

> Has everyone really become so stupid now?

Not everyone. But unfortunately, yes, very many people have.

bigstrat2003

4 日前議論

Even if you want to use an LLM (which you probably shouldn't, but it's your life), you best know how to program if you want to make anything good. They suck at programming, and need a human to guide them.

bigstrat2003

4 日前議論

We most certainly are not at that point.

bigstrat2003

4 日前議論

> Kojima is precisely what happens when you stop thinking of games as actual an interactive entertainment format and start thinking of it as a "cinematic experience" instead.

And for Western devs, we have Rockstar doing that. RDR2 is a wonderful movie, but a pretty poor game. Unfortunately, they forgot what medium they were working with.

bigstrat2003

4 日前議論

No, they aren't. Not even at the one thing (programming) which they are supposed to be best at. People are always claiming "frontier models are so much better" but that is a false claim.

bigstrat2003

4 日前議論

Anthropic's claims about their own products have almost zero value as evidence. They have lied to our faces about stuff before, and will again if it drives the hype cycle which delivers them money.

bigstrat2003

4 日前議論

> This assumes that there aren't algorithmic breakthroughs which reduce training/inference costs by several OOMs.

Yes, if one must assume something it is generally fair to assume that things will continue as they are. Research breakthroughs do happen, but they are not something for which you can predict the timing.

bigstrat2003

4 日前議論

It's a pejorative only because determinism is what makes computers useful in the first place. You get a consistent result, every single time, unlike if you have a human in the loop. Because LLMs are stochastic, they have removed the thing that makes computers useful to us, thus it's a pejorative.

bigstrat2003

5 日前議論

Not really. I didn't use GPT-2, but I don't think there was much difference between we got what out of GPT-3.5 and 5.5. It's still an unreliable tool which will go off the rails the second you aren't watching it like a hawk. It still has zero intelligence or ability to reason as to why its output is flawed. Just throwing more compute hasn't gotten us anything worth using.

コメント