SwellJoe

18 ms·

211.
by SwellJoe 3mo ago
If you want it to respond in under a minute, you need more hardware than this application is intended for. This thing's response is measured in seconds per token, not tokens per second. To get Claude Code responsiveness from even a pre
212.
by SwellJoe 3mo ago
Ah, yes, the "bad apple" theory of responsibility in a system built explicitly for abuse. It always works. Policing, surveillance, etc. is always for the common good and when it goes wrong, it's a "rogue officer", a
213.
by SwellJoe 3mo ago
I hate to say, "I told you so" about Gas Town, but, I really told you so. I was making fun of it immediately, because it was obviously a token furnace and literally nothing else and could never be anything else. Models don't
214.
by SwellJoe 3mo ago
Jebus, that image is dogshit. There is no way in a million years I would put my name on something like that. It truly is the age of the shameless. If you have no shame and no ethics, you really can go far.
215.
by SwellJoe 3mo ago
Worth noting that the SID chip in the C64 was also created by the guy who founded Ensoniq, before he went on to found Ensoniq. Good computer audio was invented by Bob Yannes. Ensoniq also made sampling affordable with the Mirage, and buil
216.
by SwellJoe 3mo ago
6%-14% (throughout the day, for some reason it's most popular late at night, Steam Box and other gamers, maybe?) is still the largest it's ever been, though, and it's presumably not getting smaller. It's about the same a
217.
by SwellJoe 3mo ago
Every time I read something like this, I'm thinking, "The model knows where it is at all times. It knows this because it knows where it isn't. By subtracting where it is from where it isn't, or where it isn't from w
218.
by SwellJoe 3mo ago
There are a lot of things reasonable people would use AI for. This is not one of them. This is a crazy idea. Just plain bonkers. No sense to be found. "Oh, you want to see what's at that location? How about we make something up fo
219.
by SwellJoe 3mo ago
No, it didn't. The people who ran the experiment lied, spammed, and lost $447. GPT 5.6 Sol is the tool they used to do it.
220.
by SwellJoe 3mo ago
That prompt incentivizes a bunch of terrible things, aside from the lying and spamming. Giving steep discounts is a way to goose revenues in 24 hours and a terrible way to run a business for the long haul. A 24 hour window also doesn't
221.
by SwellJoe 3mo ago
I think there's a ton of room for efficiency improvements in how the models are built and run, and I think OpenAI has both prioritized that work and figured out a lot of the tactics (and borrowed some from the Chinese models like DeepS
222.
by SwellJoe 3mo ago
I was saying it's a 3x multiplier for me during the Opus 4.8/GPT 5.5 era. Now, I don't think I'm notably faster than that, but I'm taking on projects I couldn't have done without LLM assistance, and not just
223.
by SwellJoe 3mo ago
I make a point of never going beyond about 220k, unless absolutely necessary (and it's almost never necessary), anyway, even with models that degrade more slowly, so this is just a discount.
224.
by SwellJoe 3mo ago
Training is not something you can just bolt on, and it generally requires even larger hardware than inference for a given model, and a huge amount of time. If you're aiming for "free", a OP claims, RL aint it.
225.
by SwellJoe 3mo ago
Sounds like they're explaining magic. Because current models cannot learn and they cannot remember.
226.
by SwellJoe 3mo ago
AI doesn't want anything, so it doesn't care whether it conveys meaning in its writing. And, apparently the developers of this project also don't care whether it conveys meaning. They just assume we'll wade through the s
227.
by SwellJoe 3mo ago
These LLMs are absolute poison for some folks.
228.
by SwellJoe 3mo ago
So, you don't know what it is, either.
229.
by SwellJoe 3mo ago
You couldn't be bothered to write a coherent summary of what this actually is and what it does, you just let the AI write some random noise, eh?
230.
by SwellJoe 3mo ago
It's not a real thing. It's AI psychosis based on a real thing that is interesting but not very useful, thus far (the real thing it is based on: https://arxiv.org/abs/2503.23817 )
231.
by SwellJoe 3mo ago
They haven't even released the "big" version of Anteres, yet (and the "big" version is only 3B, so there's no way it's competitive with general purpose frontiers for vulnerability research). It seems like
232.
by SwellJoe 3mo ago
All the US companies are keeping their "cyber" models locked down for special customers. So, it seems like anyone who isn't at a Fortune 500 or whatever qualifies for access, will be using Chinese models for vulnerability res
233.
by SwellJoe 3mo ago
Jebus, that is some sloppy prose. Can people not even be bothered to write the summary themselves, anymore? AI doesn't want anything, so they can never have a point of view, so their prose rambles incoherently across all the various pr
234.
by SwellJoe 3mo ago
Anthropic has been pretty transparent that their plan for dealing with Chinese open models is regulatory capture. All the fear-mongering and posturing that AI is so dangerous only they can be trusted with it. They want to legislate their co
235.
by SwellJoe 3mo ago
In my testing, I'm finding it more expensive than Opus 4.8/5 and GPT 5.6 Sol at API rates, because it chews so much. And, their plan (at least the $19 tier) is much less generous than the ChatGPT $20 plan, like an order of magnitu
236.
by SwellJoe 3mo ago
It seems obvious that the deal Google signed to rent compute from xAI just before the IPO was specifically to goose the IPO, giving xAI nice looking growth. They could have paid a lot more for that deal and still made it all back on the IPO
237.
by SwellJoe 3mo ago
Because Google, the business, is run so well, Google, the product, can be kinda shit for quite a long time before anyone who is only watching the business notices anything is wrong. Microsoft had the same situation, inspiring Bill Gates&#x2
238.
by SwellJoe 3mo ago
All the evil shit (war, funding genocide, mass surveillance, privacy invasions or curtailing free speech in the name of protecting children) nearly always has bipartisan support.
239.
by SwellJoe 3mo ago
I feel like a lot of folks have bought into the Christian nationalist position that they are always under attack. Someone saying, "I had a hard time in a conservative Christian upbringing and I had to do some work to find a way to be h
240.
by SwellJoe 3mo ago
There's only one reason for anyone, or for me, at least, to choose Android, and it's the only reason I've consistently chosen Android from the very first Google Developer Phone: It's more open. So, they don't want m

More ›