zarzavat

6 ms·

31.
by zarzavat 2mo ago
I thought 'how bad can it be?', yet only three paragraphs in "a specifier now names a repository rather than choosing a transport" If you're going to serve us slop then at least serve us good slop, give it a pas
32.
by zarzavat 2mo ago
Even Artificial Analysis has Opus 5 better than Fable in their aggregated "Intelligence Index" which combines 9 benchmarks. Opus 5 is heavily benchmaxxed.
33.
by zarzavat 2mo ago
It's China. It's a given that they use your data for training. At least they're nice enough to be honest about it.
34.
by zarzavat 2mo ago
3. Deployment problems unrelated to the weights causing degraded performance
35.
by zarzavat 2mo ago
Freudian slip.
36.
by zarzavat 2mo ago
The critical difference for me is if they ask while you're eating , perhaps 5 minutes after serving your table, that's OK it means they are genuinely interested in improving your experience and it's an opportunity to correct
37.
by zarzavat 2mo ago
I could say the same about US restaurants. I want to socialise with the people I'm sharing the meal with, not some random waiter. The servers are constantly interrupting, ruining the mood and you feel unwelcome like they want to get yo
38.
by zarzavat 2mo ago
It does it quite consistently on higher reasoning. I believe they have made display answer a tool call so that it can keep updating its answer as it thinks, instead of doing all the reasoning up front, essentially inverting the usual flow
39.
by zarzavat 2mo ago
Fable is indeed larger than Sol. OpenAI is developing Astra which will be more of a Fable-sized model. If you can train a larger model then you can distill smaller models from it. You don't need to necessarily serve the larger model pu
40.
by zarzavat 2mo ago
Even if it were possible it wouldn't change the outcome. China is capable of training frontier models even without distillation. Distillation is only an accelerant. The primary resource you need to train LLMs is money and China has ple
41.
by zarzavat 2mo ago
It is possible to ignore laws if it's too much bother to comply with them... it's hard to imagine the authorities will come after someone for mailing out 10 anti-static bags.
42.
by zarzavat 2mo ago
ChatGPT does this. It will present an initial response, then change its mind partway through the thinking process and present a different response.
43.
by zarzavat 2mo ago
I'm not saying that Anthropic and OpenAI are definitely dishonest, there is no proof of that. However, when you have 1) an extremely high financial incentive to be dishonest (trillions of dollars), 2) low-ish chance of being caught, es
44.
by zarzavat 2mo ago
Yes your Chinese open model could have a time-release backdoor, just as your Chinese vibrator could have a hidden microphone that records everything you say and transmits it to the CCP. But does it? No. What's much more likely is tha
45.
by zarzavat 2mo ago
You have to keep your session warm in cache. Keep the AI talking/thinking. If you have not touched a session for few minutes then /clear and start a new session. Providers will generally keep your session in cache for at least 5 m
46.
by zarzavat 2mo ago
That's effectively what it means. You can run frontier models at home e.g. Kimi K3, but you'd need a large amount of money.
47.
by zarzavat 2mo ago
You may be causing a lot of cache misses. You have to use the caching efficiently otherwise you can burn up any plan in any amount of time.
48.
by zarzavat 2mo ago
Not on Mac OS, opening more than one instance is undefined behaviour. It may work, but it may also corrupt your data.
49.
by zarzavat 2mo ago
"AI at scale" is just a euphemism for slop. Current gen AI can augment human engineers but not outright replace them.
50.
by zarzavat 2mo ago
We can tell OP is using local or unencrypted DNS because they got DNS-redirected to a message. If you have DoH configured at either OS or browser level then you will not see a message: the site will either work, or it will error out if the
51.
by zarzavat 2mo ago
It may be nostalgia but I feel like we reached peak humanness with GPT-4o and since then it's been getting more and more alien. Particularly Fable.
52.
by zarzavat 2mo ago
Unfortunately it doesn't solve the problem because malicious code can still end up compiled into the user's program which they will promptly execute, possibly in production... Supply chain attacks are not a problem that can be sol
53.
by zarzavat 2mo ago
So instead of using JS to sort the table directly, now you make a JS fetch() request to sort the table?! It does make sense in one situation, if the table is a large server-side paginated one and the sorting is really an ORDER BY clause. Bu
54.
by zarzavat 2mo ago
This statement is a little too strong. It's a mistake to care about the equality of floating point numbers after subjecting them to irrational operations. On the other hand, the entire internet runs on the fact that doubles exactly rep
55.
by zarzavat 2mo ago
> But math never decreed that sine and cosine have to take radian arguments! If you don't use radians you have to add to add conversion factors everywhere to do calculus. Radians are the natural unit for sin/cos just as E is th
56.
by zarzavat 2mo ago
Doesn't Google v Oracle provide protection here? Copying APIs is fair use. If it's patents that are the problem then presumably all these large semiconductor companies have defensive parent portfolios.
57.
by zarzavat 2mo ago
Is CUDA still a moat? Are we not at the point where frontier models can reimplement software stacks, given you throw enough tokens at the problem.
58.
by zarzavat 2mo ago
> I'd remove user-defined generics from Go, and all the overcomplicated shit related to them, including iterator functions. Go advances one blub ragequit at a time.
59.
by zarzavat 2mo ago
Models already reason in latent space, but they have to keep encoding-decoding their "thoughts" from/to human interpretable tokens, and notably those tokens are sampled from a distribution. The model cannot output a vector
60.
by zarzavat 2mo ago
There's an inverse relationship between exposure to the effects of climate change and CO2 emitted per capita. The people who will need live preserving artificial cooling the most are exactly the people contributing the least to climate

More ›