AI For Humans Show
@AIForHumansShow
AI For Humans is a podcast & YouTube channel about artificial intelligence made by humans for humans. Created, produced & hosted by @attack & @gavinpurcell
NEW EPISODE IS HERE... - @sama on @TheoVon talks GPT-5 - The White House's new AI Action Plan - @runwayml works with @netflix & @Disney And @Attack dug in and discovered some sort of very deep conspiracy that is affecting all of us? Link below 👇
New Episode! "OpenAI Teases GPT-5 as America Goes Full 'AI Action' Mode" OpenAI’s Sam Altman dribbles out GPT-5 teases as the White House’s AI Action plan lays out exactly how all-in America is on t… Player links & show notes: aiforhumans.show/e/140247752316…
This is interesting
We're launching an "AI psychiatry" team as part of interpretability efforts at Anthropic! We'll be researching phenomena like model personas, motivations, and situational awareness, and how they lead to spooky/unhinged behaviors. We're hiring - join us! job-boards.greenhouse.io/anthropic/jobs…
This is a must read report about how we don’t have enough energy to compete in AI right now but this part really got me… Single model needs 5GW?!?!
New Anthropic report: Build AI in America. We outline what it will take to ensure America has the energy and infrastructure it needs to maintain its leadership in AI.
Today, we at @OpenAI achieved a milestone that many considered years away: gold medal-level performance on the 2025 IMO with a general reasoning LLM—under the same time limits as humans, without tools. As remarkable as that sounds, it’s even more significant than the headline 🧵
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
i vibe coded a little game called Coldplay Canoodlers you're the camera operator and you have to find the CEO and HR lady canoodling 10 points every time you find them 👇link
OpenAI is testing a new model called "o3-alpha-responses-2025-07-17" on WebArena The model will appear with the name "Anonymous-Chatbot"
New Episode! "OpenAI's New ChatGPT Agent Might've Just Stolen Your Job" ChatGPT Agent is OpenAI’s new combo of Deep Research and its Operator web-browser agents into one near-human level worker. Is … Player links & show notes: aiforhumans.show/e/140199251631…
We have graded the results of @OpenAI's evaluation on FrontierMath Tier 1–3 questions, and found a 27% (± 3%) performance. ChatGPT agent is a new model fine-tuned for agentic tasks, equipped with text/GUI browser tools and native terminal access. 🧵
Something to watch out for when evaluating tool-using agents: they can "cheat" by browsing the web and simply looking up the answer key. The @OpenAI ChatGPT Agent team had to take special care to mitigate this risk.
ChatGPT agent’s capabilities are reflected in its state-of-the-art performance on academic and real-world task evaluations, like data modeling, spreadsheet editing, and investment banking.
Pro users get 400 Agent credits a month. Plus/Team get 40 Agent credits a month. Source: Mr Altman live on stream just now.
"we also have layers of monitors that peer over the agent's shoulder and watch it as it's going..."
