Loading activity…
@sriramk
ex-white house, a16z - http://sriramk.com
AI
Satya Nadella argues organizations should control what models can access and do, with safeguards outside the models themselves.
AI
Igor Babuschkin claims fine-tuned open-weight models can beat GPT-6 Astra and Claude Opus 5.5 for under 1% of the cost.
AI
Microsoft said GitHub Copilot will soon route some tasks to local models on Windows and others to cloud-scale models. The company tied that plan to its broader “hybrid intelligence” push, while separately saying MXC sandboxing is now generally available on Windows 11.
YouTubeAI
National Compute says .edu, .mil and .gov email holders can sign in, while others can join a waitlist.
AI
A poster says parts of the US government need technical expertise, especially around AI, and offers to make introductions.
AI
On X, Sriram Krishnan said a side effect of math work solved by leading models was exposure to Lean for formal proofs. He added that he had not realized how broadly Lean had been adopted in the math community over the last decade.
AI
Reflection says Beam uses 23B active parameters and plans to publish its full weights later in October 2026.
AI
One user wants fine-grained security, clearer views of AI activity and independent agents to check AI results.
AI
One observer says service integration and small UX touches matter more to them than the model powering an agent.
AI
Future macOS controls will require very explicit user action before an app receives broad system access.
AI
A participant says an Uncapped conversation with a16z growth investing head David George considers what could keep the AI buildout going or slow it.
AI
World Labs says Fei-Fei Li will join AMD as executive vice president and chief scientist, working directly with CEO Lisa Su. The company says the move will bring its spatial AI research closer to the hardware.
AI
A September 25 post says Musk plans the new set of GPUs for Colossus 2 by the end of 2026.
AI
Derek Thompson’s article examines China’s view of AI, framing the topic around seven big questions.
AI
A user says Carmack was the only name that came to mind for someone seen globally as technically, ideologically and financially neutral and trusted.
AI
The user predicts an OpenAI or Anthropic iMessage agent would undercut Instinct, citing access to existing personal context. A reply suggests asking AI to compile that context for a new assistant.
AI
One user reports about 75 inquiries for fall 2027 PhD admissions and says they’re long past being able to open all their mail. Their concern: losing a way to reach people without an introduction.
AI
One post connects the question of knowledge’s usefulness to the step-by-step automation of knowledge production.
AI
One user contrasts decades of damaging internet malware with what they describe as relatively few significant AI security incidents.
AI
The post contrasts the idea that AI “should give us back time” with its author’s enthusiasm for what they call “the intelligence revolution.”
