miroljub

8 ms·

121.
by miroljub 5mo ago
> Contrary to expectations, impolite prompts consistently outperformed polite ones, with accuracy ranging from 80.8% for Very Polite prompts to 84.8% for Very Rude prompts. These findings differ from earlier studies that associated ruden
122.
by miroljub 5mo ago
Did we already reach wokeness level five where we worry about offending a software?
123.
by miroljub 5mo ago
Yep, and then they let US companies handle that data. One more proof EU regime is run by ... no, I won't tell, don't wanna get arrested.
124.
by miroljub 5mo ago
Quick reminder that US data protection doesn't apply to non US customers. Companies are not even allowed to disclose their spying.
125.
by miroljub 5mo ago
Why should I use it instead of, for example, Brave?
126.
by miroljub 5mo ago
If the tokens are already burned, it's better for the environment to share the result for everyone to use it.
127.
by miroljub 5mo ago
You're absolutely right! My mistake. I should have said: "I like starwberries, and driving to the car wash to clean my car."
128.
by miroljub 5mo ago
I think we reached the point where it's not reasonable to expect that Microsoft is capable of improving anything. Their products all go in the same direction. The best case for Microsoft software I hope for is that they do only some mi
129.
by miroljub 5mo ago
Of course Anthropic is doing the same. Just ask Claude in Chinese to introduce himself and tell you what he does and how it can help you. There's a good chance that it will tell you its name is DeepSeek ;)
130.
by miroljub 5mo ago
> I like Claude a lot. I hope it continues to be good I like strawberries. And car washing. I hope they stay unaffected by Claude's whims.
131.
by miroljub 5mo ago
There is no subsidy in interference. Serving existing models is a highly profitable business. The proofs are many independent companies serving models for profit. What is expensive is model development and training, but again, that's n
132.
by miroljub 5mo ago
How is that an issue? Many funds that replicate an index are not doing full replication anyway. They do some kind of sampling that includes way fewer stocks than defined in the index. What would stop a fund from just not including those sto
133.
by miroljub 5mo ago
> Do you have a reason to believe the gap between AI (not LLMs specifically, but AI generally) and human intelligence will peak near the difference between human competitors (what... 20-30 IQ points)? So we established that competing hum
134.
by miroljub 5mo ago
"Something" is in use.
135.
by miroljub 5mo ago
If we look at our human history, there are millions of examples where less intelligent beings destroyed highly advanced civilizations. It was never about intelligence, but about willingness to destroy (willingness to defend is not enough).
136.
by miroljub 5mo ago
My guess is they are still in the "fake it till you make it" phase. There's no Mythos, it's just a hype machine fueled by a hot air.
137.
by miroljub 5mo ago
It's pretty arrogant for a foreigner to tell me how to spell his county name in my own language, which doesn't even have those letters in an alphabet. All while adding -stan suffix to my country name in their language. In short, i
138.
by miroljub 5mo ago
EU is totalitarian. If Europeans want to have a chance of protecting what's left of their freedoms, there's no other way than dissolving EU completely. It's not possible to reform that monster.
139.
by miroljub 6mo ago
Editing config files is not necessary. Just do /login from your session, choose your provider, and there you go.
140.
by miroljub 6mo ago
This one still lacks some features. They still recommend using their OpenAI compatible endpoint. But I guess Anthropic is just not capable of implementing the OpenAI API compatible client in Claude Code.
141.
by miroljub 6mo ago
> even when model subsidization is done, those open source models are quite viable alternatives. Model inference was never subsidized. Inference is highly profitable with today's prices. That's why you have many inference provi
142.
by miroljub 6mo ago
I wonder why the question about data security and training comes often with DeepSeek, Kimi, Glm and never with Anthropic, OpenAI, and Google models. Why is that? IIRC, USA data protection protects data of US citizens only, foreigners data i
143.
by miroljub 6mo ago
Even worse, they exercise bias on a daily basis.
144.
by miroljub 6mo ago
I don't have an opinion about Zig AI policy for contributions. Their project, their policies. Fine for me. However, I wanted to give Zig a try in an agentic coding scenario. For tasks that would take a few seconds when choosing Python,
145.
by miroljub 6mo ago
SourceHut never really clicked for me. It doesn't give me anything useful that I don't already have in a bare git repo through a ssh. Forgejo, on the other hand, is a drop-in replacement for GitHub.
146.
by miroljub 6mo ago
> Anyone who has ever used Mercurial knows very well what a good versioning tool UX looks like... So true. I used Mercurial back in the day and also used Darcs before it, and it helped me realize that the best versioning tool UX that exi
147.
by miroljub 6mo ago
This subsidized inference is just a marketing ploy to increase prices and profit. If common people can have a DIY setup with an open source model cheaper than those behemoths with a scale advantage, it's clear that we have been playe
148.
by miroljub 6mo ago
It's his opinion based on observing his social group. Should be clear from context.
149.
by miroljub 6mo ago
Foreign state existence.
150.
by miroljub 6mo ago
No. Just a notice that above mentioned group frequently cry for source when they don't agree with an opinion. Opinion is personal, it's not a fact and doesn't need source.

More ›