Tootfinder

Opt-in global Mastodon full text search. Join the index!

@ErikJonker@mastodon.social
2025-08-06 12:22:31

Open Weight language models are released by OpenAI.
Interesting what the experiences will be on local configurations , 16GB (V)RAM is a lot but attainable for a lot of people.
#openai

@tinoeberl@mastodon.online
2025-06-10 06:08:23

Große KI-Unternehmen wie #Anthropic und #OpenAI geraten zunehmend in Konflikt mit erfolgreichen Apps, die ihre Modelle nutzen.
Die populären Anwendungen #Windsurf und

@stefanlaser@social.tchncs.de
2025-08-12 05:53:23

#ChatGPT5, it appears, is full of shit.
"#OpenAI’s products are no longer primarily aimed at consumers but at investors. As long as you avoid a full-scale user revolt (which GPT-5 actually did incur…), you can continue to assuage or even attract more backers on your path of relentless e…

@ErikJonker@mastodon.social
2025-08-08 07:42:58

The graph error OpenAI made has nothing do with AI but with internal organisation and quality controls. In my own work it would be unthinkable that an error of this magnitude would slip through to the live presentation. How can/could this happen, is OpenAI a very strange / atypical organisation?
And if they can't control the quality here what means that in a general sense?
#OpenAI

Graph with faults shown during OpenAI live presentation of GPT5
@davidshq@hachyderm.io
2025-06-05 01:34:09

How did I just learn today that #OpenAI is retaining all #ChatGPT chat logs due to a court order: linkedin.com/posts/davemackey_

@bici@mastodon.social
2025-06-08 23:35:58

Sam & Jony
Sam & Jony was sweethearts, oh LoveFrom how they did love
Swore to be true to each other, true as OpenAI above
Jony was his io, he wouldn't do Sam wrong

Sam & Jony were lovers
#IO #OpenAI

@michabbb@social.vivaldi.net
2025-06-05 09:16:12

💼 #OpenAI actively hiring #Rust developers for #codex-cli team expansion

@grumpybozo@toad.social
2025-06-07 23:00:49

This might also explain why #OpenAI pushed ChatGPT to market before it was worth touching. If you can’t afford testing at scale, why not do a huge free public release so that you can call the last 80% of the dev effort “product support" which should be fully expensible.
NOTE: I AM NOT A TAX EXPERT AND THE ABOVE COULD BE ABSURD.

@mho@social.heise.de
2025-05-28 20:36:40

"In Empire of #AI, journalist Karen Hao writes about the rise of #OpenAI and the impacts of AI around the world. Below is an extract from the book on the effects on Chile's mineral reserves and water resources."

@laimis@mstdn.social
2025-06-19 14:09:48

This seems like a cry for help from a few remaining honest individuals at #openai:
openaifiles.org/
This is why I do not use their LLMs, regardless how powerful their models might be. W…

@thomasrenkert@hcommons.social
2025-05-23 08:15:29

The #OpenAI paper by Baker et al, "Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation" comes to a troubling conclusion: #LLM s with #reasoning or

If CoT pressures are used to improve agent capabilities or alignment, there may be no alternative approach to yield the same improvements without degrading monitorability. In the worst case, where the agent learns to fully obscure its intent in its CoT, we ultimately revert to the same model safety conditions that existed prior to the emergence of reasoning models and must rely on monitoring activations, monitoring potentially adversarial CoTs and outputs, or improved alignment methods. Model a…
@mho@social.heise.de
2025-08-13 18:40:04

Das #GPT5 keine Karten zeichnen kann, ist übrigens Absicht.
Jetzt wisst ihr Bescheid. 👇🏼
#ChatGPT #OpenAI #SamAltman


Viele wollen OpenAI fallen sehen

Mein educated guess ist, dass die vielen Fehlleistungen sogar absichtsvoll in Kauf genommen worden sein könnten: Eine riesige Zahl von Leuten macht aus schierer Häme nichts anderes, als Schwachstellen von GPT-5 zu finden und diese zu veröffentlichen. Eine schnellere, umfassendere und bessere Fehleranalyse des eigenen Produkts ist kaum zu bekommen. Und mit genau diesen Daten kann die Realtime-Router-KI wiederum rasch lernen, besser zu werden. Das ist nur eine V…
@digitalnaiv@mastodon.social
2025-06-20 06:23:10

#OpenAI und #Microsoft – eine Beziehung wie im Hollywood-Drama: Liebe, Macht, Milliarden. Während die beiden sich um geistiges Eigentum streiten, winkt das Pentagon mit einem fetten Scheck. Wer kriegt die Kontrolle? Klingt nach einem schlechten Thriller – nur leider ist es Realität. Wer braucht sch…

@michabbb@social.vivaldi.net
2025-08-14 09:30:10

🎵 #FFmpeg 8.0 integrates #Whisper for local audio transcription 🎵 #ai
🔧 Direct integration of #OpenAI

@michabbb@social.vivaldi.net
2025-06-05 09:16:11

#OpenAI #Codex CLI Goes Native with #Rust Rewrite 🚀
⚡ Rewritten from #TypeScript to

@davidshq@hachyderm.io
2025-06-23 21:59:20

The projects functionality in chatgpt shows promise but lacks some basic functionality - e.g. sorting, reordering, nesting.
Is anyone aware of code out there to manage / override the projects UI?
I've done some basic vibe coding but haven't had great success.
Strange to me, but they only show a few projects at a time, you have to paginate through them and the API grunts about auth.
I see some projects on github, etc. but I'm not seeing anything focused on projects particularly.
#question #ai #llm #projects #chatgpt #vibecoding #openai

@michabbb@social.vivaldi.net
2025-07-28 19:38:17

#GLM45 frontier #AI #LLM with 355B parameters ranks 3rd against #OpenAI