Bro digitally simps for nations for political identity, and instantly folded under the lightest application of reasoning. That is sad.
360 karma · joined December 21, 2023
Bro digitally simps for nations for political identity, and instantly folded under the lightest application of reasoning. That is sad.
Really? How do you even know that? You think another round of price hikes within the year is unimaginable, which what the economic consensus on immediate tariffs this high predicts?
The unemployment rate is 4%. The amount of liberation day tariff supporters is an order of magnitude higher than that. Pretending that things can't get worse is dangerous and stupid.
The post has a catchy title and a (in my opinion) clear message about using models as API callers and fuzzy interfaces in production instead of as complex program simulators. It's not about using models to write code.
Social media upvotes are less frustrating imo if you see it as a measurement of attention, not a funneling of value. Yes people like things that give them value but they also like reading things with a good title.
Radioactive isotopes in banana: consume away.
Understand now? I'm not even defending anything other than how two things that vaguely appear contradictory, are in fact, not.
Would you elaborate on what you mean by friction with micro transactions?
[0] sure this still applies, but sponsors and Patreons exist at a far higher rate than in 2015 when I first thought this, and it's on Google/youtube to figure out their monetization policy, not me.
That aside, when I read about downsides of universal healthcare in America, I still think the tradeoff is worth it. Even if it means wait lines and decades slower drug discovery, if it meant measurably less rage at the system, I would support it. I know this isn't the thread topic, but the public polled reactions to literal assassination being not particularly negative points to how unpopular the current system is, despite other measures showing it as popular.
Anyway, screw the (political) Kennedy family. Seriously.
If you use speculative decoding (a small model generates tokens verified by a larger model, I'm not sure on the specifics) you can get past 20 tokens per second it seems. You can also fit 32B models like Qwen/Qwen Coder at Q6 with lots of context this way, with spec decoding, closer to 40+ tks/s.
Obviously the drop in cost for capability in the last 2 years is big, but I'd wager it's closer to 10x than 100x.
I also think this argument means that theoretically someone with a 6th grade education and a 12th grade education should have the same economic output, as they both know how to read and add/multiply numbers. In real life the reason someone drops out before/during high school is always extenuating life circumstances, so the outcomes are much worse for them. But in a world where that didn't happen, do you think the outcomes would be essentially identical?
The part about this which is weird is it's not like GPT would have access to information about him that's not already public on the web. If this is a deliberate censorship PR decision, it's incredibly stupid and backfiring. What else explains this behavior?
[1] https://www.theguardian.com/world/2018/dec/16/akhmed-one-arm...