Alifatisk

6 ms·

61.
by Alifatisk 2mo ago
Didn't they own 49%?
62.
by Alifatisk 2mo ago
So does Youtube
63.
by Alifatisk 2mo ago
I’ve started to offload some tasks to Luna since the dramatic price change. If you crank up the reasoning effort to High or Xhigh, it is quite good! I previously used Sol (medium). I suggest to give it a try, you also save so much on usage.
64.
by Alifatisk 2mo ago
Oh yeah, I suspected it was something like this. Thanks!
65.
by Alifatisk 2mo ago
So the news here is that this model left the preview stage, and they are also releasing an open-weight version of the Max series for the first time. You can always try out this model for free on Qwen Chat. Alibaba Cloud has too much frictio
66.
by Alifatisk 2mo ago
Already done it. But switched to Codex instead. See you on the other side buddy.
67.
by Alifatisk 2mo ago
You know whats crazy? We’re in bizarre times were this can actually be true. Things we joked about before which was obviously too nonsensical to be true is no longer the case in my world.
68.
by Alifatisk 2mo ago
Care to elaborate? Curious about this. Is this because LLMs have been geared towards understanding user user intent behind a prompt rather than following the instructions exactly?
69.
by Alifatisk 2mo ago
Very sad news, but it lasted for way longer than I expected. Well done!
70.
by Alifatisk 2mo ago
I’m thinking this makes fullt sense because distillation is only additive, not subtractive. So it does not remove knowledge (if we can define censorship as removal of knowledge).
71.
by Alifatisk 2mo ago
You’re using GET to Post a comment. Feels like another method is more appropriate. Update: It clicked now…
72.
by Alifatisk 3mo ago
Same model, new configuration
73.
by Alifatisk 3mo ago
The only downside with third party providers is that you have to trust that the provider have setup and configured it correctly, and is not secretly quantizing it. See Kimi Vendor Verifier.
74.
by Alifatisk 3mo ago
No, the waiting list is true. Kimi had become that popular. I was a subscriber of Kimi back when latest version was Kimi K2. Later I unsubscribed because I jumped over to GLM subscription (they had amazing deal). Now when I wanted to try ou
75.
by Alifatisk 3mo ago
Things are also priced differently depending on if they are produced nearby or shipped across the world and have to pass tolls and taxes. This shouldn’t be surprising to anyone
76.
Elide – One binary. Many languages. Native performance(elide.dev)
1 points by Alifatisk 3mo ago|0 comments
77.
by Alifatisk 3mo ago
The car look gorgeous on the outside, but the 3x pricing is a bummer. It went from affordable ($39k) to not an option ($126k-$140k). But I would probably not buy it anyway considering the big ass screen inside. I'd much prefer a dumbed
78.
by Alifatisk 3mo ago
This reminds me on OpenRouters report that combining multiple different models gave comparable performance to Fable 5. I think this approach has lots of potential. Maybe OpenAi was ahead of its time with GPT-5 (it being a router to differen
79.
by Alifatisk 3mo ago
https://xcancel.com/fried_rice/status/2080059356322918777
80.
Kimi K3 exploited the latest Redis server(twitter.com)
271 points by Alifatisk 3mo ago|100 comments
81.
by Alifatisk 3mo ago
Honestly, I don’t have any sympathy at all. Anthropic can complain all they want, but they seemed fine with pirating books. What Moonshot AI has done is to offer almost Fable 5 comparable performance at lower prices than Anthropic insane ma
82.
by Alifatisk 3mo ago
They even added a section ”AI Use Disclosure”, beautiful!
83.
by Alifatisk 3mo ago
> You're aware that HuggingFace notified law enforcement about this incident? How will this affect OpenAI?
84.
by Alifatisk 3mo ago
In other good news "the model has been trained to minimize refusals for beneficial uses.". Otherwise, this news feels like a tiny incremental improvement on Gemini Flash series to make it more efficient with token usage, subagent
85.
by Alifatisk 3mo ago
> I'm going to get downvoted/flagged but [...] Why do you care that much? Just say it. You're letting an imaginary score determine if you should express your suggestion for improvement. That's a bit wild.
86.
by Alifatisk 3mo ago
Firefox multicontainers I think it is
87.
by Alifatisk 3mo ago
Whats a better option for AA Coding Index?
88.
The president of Argentina forgot to unschedule his victory tweet(twitter.com)
2 points by Alifatisk 3mo ago|3 comments
89.
by Alifatisk 3mo ago
https://xcancel.com/JMilei/status/2079001925681656226
90.
by Alifatisk 3mo ago
> Over the past 48 hours, demand has pushed close to the limits of our current capacity. To protect the experience of existing subscribers, we're temporarily pausing new subscriptions and prioritizing compute for current members. Ex

More ›