121–130 of 146 posts
Earlier quoted context omitted.
Censorship is the suppression of speech, public communication, or other information. (Wikipedia)
and which part do you think Facebook/twitter/etc don't take part in?
Earlier quoted context omitted.
I have no trouble with Google as they have a way less aggressive crawl rate and Google actually sends me website traffic. Claude only takes. Google takes and gives.
In the example above, Claude's crawler is on a whooping 0.0008 TPS, not exactly a DDOS attack. I'm not against blocking Claude for the reason you outline (or all LLM crawlers for that matter), but it's not exactly an "aggressive crawl rate".
Earlier quoted context omitted.
the chilling effect is the same.
No, the chilling effect is very different. One is a tragedy that needs to be brought to justice, the second one is never going to get any semblance of justice from its very principle.
Earlier quoted context omitted.
and which part do you think Facebook/twitter/etc don't take part in?
Wrong question. Of course they censor content. I did not say they don’t do it, I only pointed out that comparison to USSR is quite an exaggeration.
So, in general Claude still beats newest GPT-4o model in software development category, right? [1] Although new GPT is better at small coding excercises (72.9% vs 68.4%), it falls behind in more challenging refactoring tasks (62.9% vs 72.3%). I used Claude through VPN from Europe, and from the start could see it was much better than GPT-4, in my limited very subjective use cases. [1] https://aider.chat/docs/leaderboa…
> So, in general Claude still beats newest GPT-4o model in software development category, right? [1] I'm genuinely curious to which extent widely published benchmarks can be used for this kind of assessment. The exercises this benchmark uses seem to be available in https://github.com/exercism/python/tree/main/exercises , and it seems that there are solutions at least at https://exercism.org/tracks/python/exercises/se…
I agree that it might not be fair. However, refactoring is a different story imo since as I see it relates to more general capability of the model (description about refactoring tasks are in the link from previous comment).
Earlier quoted context omitted.
No, the chilling effect is very different. One is a tragedy that needs to be brought to justice, the second one is never going to get any semblance of justice from its very principle.
tell that to the school teacher from Birmingham that is in hiding because they showed a cartoon to children. I'm sure they're real happy about justice.
Earlier quoted context omitted.
Wrong question. Of course they censor content. I did not say they don’t do it, I only pointed out that comparison to USSR is quite an exaggeration.
what evidence would convince you that the comparison is apt?
The voice user interface for natural two-way conversation is GPT’s killer feature. Gemini also suppports voice input, which I use to make interactions easier on desktop. Hopefully Anthopic will add at least speech-to-text support, although I understand smartphones often include that already. That said, the OpenAI Whisper TTS is really nice for English language input.
I find the two way voice to be unnatural and difficult to manage. I also find myself needing more time to compose my thoughts in the middle of a prompt than the ChatGPT conversation input allows me. The weird breaths and stuff kinda creep me out.
I still believe that voice user interface is one of the more power and profound shifts these tools offer. They’re among the first generation of technologies that can parse human speech, follow commands, and if desired respond with speech. Those capabilities are more powerful when coupled with the ability to compress vast amounts of knowledge and generalize.
Like anything, LLMs and multi-modal models are far from perfect and raise many ethical and philosophical concerns. I’m cautiously optimistic as I believe voice interaction can make technology much more widely accessible and helpful to many people.