alvis

7 ms·

1.
GPT-6 Astra System Card [pdf](deploymentsafety.openai.com)
4 points by alvis 13d ago|0 comments
2.
by alvis 13d ago
"Once it is available in the API, Astra will cost $10 per million input tokens and $50 per million output tokens. That is 2.5 times Sol’s current promotional price, although it matches Anthropic’s pricing for Fable 5.1." Open AI f
3.
OpenAI launches GPT-6 Astra and says welcome to the "AGI era"(thenewstack.io)
6 points by alvis 13d ago|1 comments
4.
Fable 5.1 System Card(anthropic.com)
16 points by alvis 15d ago|1 comments
5.
A Server Lost Power at 00:32. We Found Out at 08:18(danubedata.ro)
10 points by alvis 24d ago|9 comments
6.
The Maintainer Is the Interface(kennethreitz.org)
3 points by alvis 27d ago|0 comments
7.
MX Records: The Untold Story of Email(blog.lacnic.net)
2 points by alvis 28d ago|0 comments
8.
Banksy works cost public almost £150k(bbc.co.uk)
26 points by alvis 1mo ago|85 comments
9.
Google DeepMind CEO Demis Hassabis steps down(ft.com)
3 points by alvis 1mo ago|1 comments
10.
What went wrong with data lakes? A 15-year reality check(arxiv.org)
54 points by alvis 1mo ago|16 comments
11.
JSON Is Lying to You(blog.gaborkoos.com)
6 points by alvis 2mo ago|1 comments
12.
by alvis 2mo ago
Basically lunar at extra level can cover all use cases scenarios other than those requiring opus up. Goodbye sonnet and haiku
13.
by alvis 2mo ago
These models don't take claude.md, Claude Code does, and claude.md still get fed to these models if you're using Claude Code. That said, prompting on opus 5 needs new technique: https://platform.claude.com/docs&#x2
14.
Prompting Claude Opus 5(platform.claude.com)
3 points by alvis 2mo ago|0 comments
15.
Mid-Conversation System Messages(platform.claude.com)
3 points by alvis 2mo ago|0 comments
16.
by alvis 2mo ago
Also the cost per task. It appears to be significantly cheaper, cheaper than sonnet!
17.
by alvis 2mo ago
What really impress me is opus 5 is better in alignment than fable 5!
18.
Claude Opus 5(anthropic.com)
1778 points by alvis 2mo ago|1334 comments
19.
by alvis 2mo ago
here we go
20.
Claude Voice Mode(twitter.com)
3 points by alvis 2mo ago|1 comments
21.
ChatGPT Voice is now in the desktop app(twitter.com)
3 points by alvis 2mo ago|0 comments
22.
GPT‑Red: Unlocking Self-Improvement for Robustness(openai.com)
35 points by alvis 2mo ago|1 comments
23.
by alvis 2mo ago
It was almost from no more access to frontier models to having too much access in the past hours
24.
by alvis 2mo ago
Who would expect fable will really get removed given the competition? But 50% more limit is completely unexpected to me
25.
by alvis 2mo ago
I'd say 5.6 Sol performs between opus and fable, but not yet fable. Just that given its cost efficiency, claude subscription isn't attractive at all when fable is removed.
26.
Claude Code May–July 2026 weekly limits promotion(support.claude.com)
46 points by alvis 2mo ago|66 comments
27.
by alvis 3mo ago
Ironically, the key message of today's release is that Sonnet 5 is far less capable than Opus 4.8 and Mythos 5. It's a funny development is the past few weeks
28.
Claude Sonnet 5 is here(twitter.com)
3 points by alvis 3mo ago|0 comments
29.
by alvis 3mo ago
What I starting to hate is that each model's effort level can mean completely different power. Today sonnet 5's med level effort is equivalent to sonnet 4.6 low level effort :/
30.
Claude Agents in Notion(notion.com)
2 points by alvis 3mo ago|0 comments

More ›