571–580 of 594 posts
Earlier quoted context omitted.
Considering the fact that Google/Anthropic/OpenAI have WAY more compute and the race is this close, it's obvious that DeepSeek/GLM/Qwen teams are better or we're approaching a wall in terms of progress.
US gave China a gift by restricting GPU, they made them more resourceful. Too much money/resources is often a disadvantage.
Earlier quoted context omitted.
Bandwidth is really cheap in bulk. You can get a 100 gigabit internet connection for about $10,000 a month. If you were to somehow keep that saturated 24/7, you'd move about 30 petabytes in a month, so your per gigabit cost is only $0.0003. Realistically, if you had it 5% utilized, those million tokens would cost you about $0.0000062, which is pretty insignificant compared to what they charge you. (Assuming one byte…
People read the AWS rate card for bandwith and think has something to do with reality. Even though it is 1000x higher!
Earlier quoted context omitted.
The field is much older, MIRI is ~20 years old. Look up Eliezer Yudkowsky.
Yes, he's the exact reason people are distrustful. He's a crank who learned about reward hacking and made a new religious movement out of it, pretending it's a world-ending issue and deliberately avoiding much more serious issues like the concentration of power. Typical cult leader and manipulator, and his disciples in charge of major AI shops aren't any better.
If he's a cult leader, he's awfully bad at it. No opulence, no compounds, no dogma, no doctrine...
Earlier quoted context omitted.
…are you sure a brave stance against safety and welfare is what we need in this moment? Why do you think your conception of the dangers are more accurate than all the scientists who have spent their lives studying this?
> …are you sure a brave stance against safety and welfare is what we need in this moment? Is it out of convenience to not see the hypocrisy? "Safety and welfare" for you and me . Yet if you work at Anthropic or OAI, or are a partner of them then you can let it rip! Oh, and when they illegally do just that - you get a "we're sorry bro" blog post that's designed to drum up FOMO and, most importantly, zero accountabilit…
2. You're misinterpreting this as a battle over what kind of topics you can use a hosted chatbot for, and which are forbidden for corporate reasons. That is, to say least, small potatoes.
3. Blaming the companies for "zero accountability" is pretty odd. All of this is brand new, and the two big ones are both pushing for new laws on this very thing.
4. Your last point... I'm not sure I understand, sorry. They're experts in AI. Are you saying that they need to be experts in, say, bioweaponry? If so, that doesn't really follow IMO.
5. Pointing out an example of the government comissioning a private corporation to build a system to drack dissidents is exactly the "safety and welfare" work that I'm a proponent of!
Earlier quoted context omitted.
there's probably fewer bibles in China so less source material to reference I suppose
Aren't most American bibles printed in China? (Source: hazy memory)
Earlier quoted context omitted.
That’s just how Chinese words work
Depending in what you are doing, you can unlock a lot more knowledge by translating your query to Chinese and then asking that. :)
Subjectively it seemed to reason faster and better, but sometimes it would indeed output responses in Mandarin nonetheless; so I didn't pursue it further.
Earlier quoted context omitted.
Ignoring the obvious ai slop webpage. I don't really trust the benchmark. It seems either pretty saturated, or inconsistent just based on the results.
What seems inconsistent? The coverage is quite small, only 22 tests. It's more to compare the cost/speed/consistency between models, given the same tasks.
Earlier quoted context omitted.
What seems inconsistent? The coverage is quite small, only 22 tests. It's more to compare the cost/speed/consistency between models, given the same tasks.
Right. I got the feeling of it being saturated because all the top 5 fully completed it.
Also because models now are actually not that good on knowing things (domain knowledge), as they rely more on web search on tool use. So if I added a question, about some obscure fact, probably the SOTA models would fail it, but in practice they would find it with web search enabled. Not sure how to handle that. This is also why Gemini is on top, it's good enough at coding and instructions following, while having by far best general and domain specific knowledge.
Earlier quoted context omitted.
You have completely misunderstood what I was saying so badly I can't even formulate a response other than to suggest you read my reply again. I was not suggesting that LLMs are programmed with print statements, for fuck's sake.
If you say so. > When you write a program to predict tokens based on context, seeding its context with something that makes it predict "self-reflecting" text is trivial. Program does what it is programmed to do. Would observing the output of the following program inspire doubt as to its sentience? Then you follow it up with print statements as if that is a good analogy. As I said, they are not programmed, so your que…
To put it in ELI5 terms: run a program against a book, counting how many times "I love " appears in the book. Note "dogs" 4 times, "cats" 5 times, "you" 1 time into a database. Then run a program against that database. When inputting "I love" as the preceding text, the second program determines the most likely result is "cats" and returns "I love cats" (or returns "I love cats" 50% of the time, or dogs 40% of the time, or you 10% of the time, or some variation by different methods of weighting).
Yes, this is an extreme simplification. Yes, the model is not technically a database either. But this is fundamentally the process followed. You would consider it a single program if the training framework and inference engine were part of the same software and stored the computed training values to memory instead of disk, taking an input dataset and an input context as params and returning "I love cats" as the output. There's all kinds of incredibly sophisticated techniques applied on top of this foundation to vastly improve the statistical modeling and efficiency, but the underlying basics have not fundamentally changed.
> Then you follow it up with print statements as if that is a good analogy.
The print statements were not an analogy. They were pointing out the ridiculousness of doubting whether software is conscious because it generated self-referential text. Gettting software to generate self-referential text is as easy as `print(self_referential_text)`. So the only question is how the self-referential text is generated. For self-referential text generation to be more interesting than passing it as a literal print value, there would have to be some really wondrous "how" going on. But, it turns out, the "how" of an inference engine isn't that much more interesting than literally doing a `print`.