Sam Altman focused on AI cybersecurity, urging urgent collective action and encouraging organisations to use AI models defensively 110. He said OpenAI had paused some frontier reinforcement-learning training because capability gains were outpacing safety and alignment, while supporting shared industry standards and continued broad access 7. He also discussed product and infrastructure progress, including Astra’s delayed wider release, model improvements, a new chip, and business privacy 561315.
On the record
last 31 daysNo positions on the record in the last 31 days — retweets, links and asides don't count, only a view of their own.
Their output
17 in August 2026Stories it travels in
i
The stories
- ○ a voice, on the left
- ◌ a voice raising half of these
- ● a subject, on the right, under its story
- ⌒ a line is a voice raising a subject — thicker, more items; colour is the story
The stories this entity travels in: on the right, each story with the subjects seen alongside it; on the left, the voices that raised them, and a line for every subject a voice raised, bundled by story. Tap a voice to light everything it raises, a subject for who raised it, a story for its own lines. The crumb goes back.
Who brings it up
- Jack Farley 3 mentions
- Max Wiethe 2 mentions
- Warren Pies 2 mentions
- Guy Adami 1 mention
- Dan Nathan 1 mention
- Kaiser Kuo 1 mention
- Tian Yang 1 mention
- Evan Osnos 1 mention
What they talk about
- OpenAI, L.P. 6 items
- Jensen Huang 1 item
- achieving criticality 1 item
- Jason Wenk 1 item
- Stats Wire 1 item
- Oklo Inc. 1 item
- Brian Huang 1 item
- Eric Wallace 1 item
See what your own voices say about Sam Altman, every morning.
Start a desk