#claudeopus45 — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #claudeopus45, aggregated by home.social.
-
https://winbuzzer.com/2026/02/06/metr-five-hour-ai-claim-misunderstood-graph-xcxwbn/
METR's Five-Hour AI Claim: Why Everyone Misunderstood the Graph
#AI #METR #Anthropic #Claude #ClaudeOpus45 #AIResearch #AISafety #AIBenchmarks #LLMs #AICoding #AIAgents #AgenticAI #AICoding #AISafety #AIEthics
-
https://winbuzzer.com/2026/02/06/metr-five-hour-ai-claim-misunderstood-graph-xcxwbn/
METR's Five-Hour AI Claim: Why Everyone Misunderstood the Graph
#AI #METR #Anthropic #Claude #ClaudeOpus45 #AIResearch #AISafety #AIBenchmarks #LLMs #AICoding #AIAgents #AgenticAI #AICoding #AISafety #AIEthics
-
Qwen3-Max-Thinking, GPT-5.2급 추론 능력 갖춘 새 모델 공개
Alibaba Qwen 팀의 최신 추론 모델 Qwen3-Max-Thinking 공개. GPT-5.2급 성능과 자율적 도구 선택 기능으로 복잡한 추론 작업 향상. -
Qwen3-Max-Thinking, GPT-5.2급 추론 능력 갖춘 새 모델 공개
Alibaba Qwen 팀의 최신 추론 모델 Qwen3-Max-Thinking 공개. GPT-5.2급 성능과 자율적 도구 선택 기능으로 복잡한 추론 작업 향상. -
Claude Opus 4.5 is raising the bar for enterprise AI, while Anthropic’s new models dominate the coding scene. Discover how these large language models are reshaping code generation—from OpenAI’s latest tricks to Cursor’s programming AI—and why it matters for open‑source developers. #ClaudeOpus45 #Anthropic #AICoding #CodeGeneration
🔗 https://aidailypost.com/news/claude-opus-45-boosts-companies-anthropic-leads-ai-coding
-
The last 10 ways I used Claude and ChatGPT
Claude Opus 4.5:
- A discussion about this LessWrong essay on ‘parasitic AI’ with a focus on untangling the epistemology and ontology of what’s being discussed.
- A discussion about the quote from Maurice Blanchot I blogged about this morning
- A discussion about the strange mood of the film Black Bear and suggestions of other films which might scratch the same itch
- A discussion about the political economy of data centres in which I shared this blog post I wrote
- Advice drafting an e-mail I was slightly nervous about getting the tone right for
- Discussion of a call for papers in which the model reviewed a whole category of my blog to help me pull together ideas for an abstract
- Discussion of my book Platform & Agency in order to help me plan an introductory presentation about it
- A long conversation about the notion of transformational objects in Bollas in the course of which I wrote a series of blog posts I shared with the model
- A discussion of speaking invitation which I ultimately turned down and advice on how to say no in a polite way
- A discussion on AI slop in which I got the model’s feedback on some metaphors I was using (affect mining, engagement farming) I wasn’t convinced by
ChatGPT 5.2:
- A request for film recommendations following Black Bear to see if the model provided me more interesting recommendations than Opus. Interestingly, it did.
- A discussion of whether a feeling of restricted movement in my abdomen during reformer pilates should worry me given I had hernia surgery a few years ago
- A discussion of Stephen Graham’s character in the film A Patch of Fog and other roles which have a similar feel
- An update about a local business which I’m keeping track of for convuluted reasons
- Help adapting one of Maggie Archer’s diagrams for a presentation which I ultimately decided not to use
- Request to help me find a PDF of a book for a reading group which the model reliably refused, in spite of my attempt to brow beat it into finding one
- Text extraction from a badly scanned PDF which I could barely read for a course I’m on
- Testing whether the model could identify the author of a unpublished text
- Advice about gain and echo reduction with my external microphone
- Helping me finalise branding and presentation for a new network I’m launching
-
Anthropic Confirms 'Soul Document' Used to Train Claude 4.5 Opus Character
#AI #Anthropic #Claude #LLM #GenAI #AISafety #AIEthics #MachineLearning #ClaudeOpus45
-
Anthropic Confirms 'Soul Document' Used to Train Claude 4.5 Opus Character
#AI #Anthropic #Claude #LLM #GenAI #AISafety #AIEthics #MachineLearning #ClaudeOpus45
-
Claude Opus 4.5 now preserves its reasoning steps, cutting the forgetfulness that plagued long‑running prompts. Compared with GPT‑5.1 and Gemini 3 Pro, it stays on task in extended coding sessions and complex thinking blocks. Curious how this changes the LLM landscape? Read the full analysis. #ClaudeOpus45 #GPT51 #Gemini3Pro #ReasoningSteps
🔗 https://aidailypost.com/news/claude-opus-45-retains-reasoning-steps-avoiding-forgetfulness-long
-
Claude Opus 4.5 tops SWE‑bench in 7 of 8 languages, beating Sonnet 4.5 by 15%. The win spans Java, Python and other major stacks, showing Anthropic’s multilingual edge in AI‑assisted coding. Curious how the models compare? Dive into the benchmarks and see what this means for open‑source developers. #ClaudeOpus45 #Sonnet45 #SWEbench #MultilingualAI
🔗 https://aidailypost.com/news/claude-opus-45-leads-swebench-7-8-languages-15-ahead-sonnet-45
