Treasury proposes broad AI investment restrictions - Tustin AI Daily Briefing, Oct 23

Tustin AI Daily Briefing · Oct 23, 2026 · 6:56

Listen in the Podli app 🎧

Follow your favourite podcasts, listen offline and in the car with CarPlay and Android Auto, and always pick up where you left off. Free to try.

Stories in today's episode — tap a headline to read the full piece:

Treasury proposes sweeping new outbound-investment restrictions covering all AI systems
OpenAI's round is now set to close within a week at up to $1.25 trillion, led by Coatue and UAE investors
Modal's CEO on why AI compute demand isn't "rectangular" — and the procurement machine behind a $16B valuation
"You can close an eight-figure deal in two weeks" — how AI sales orgs are actually structured
81% of companies say their GPUs run at half capacity or less — and only 10% have a finance-reviewed ROI number
DeepSeek raises nearly $16 billion days after the U.S. ordered its tech off federal systems
Anthropic went from 2,500 to over 30,000 merged code changes a day — and had to abandon Git
Railway breach: Every used agents to run its own credential-rotation incident response
17 of 40 teams that claim to enforce AI agent permissions still share the credentials behind them
The Nuclear Company proposes a 6.6 GW Kentucky nuclear campus; Navy picks Radiant for a shore-based microreactor
GenReasoning: training agent organizations as collectives — and the oversight that erodes as they coordinate
Mollick: prompt injection may be a smaller problem in new models
Nate Jones: retire long-lived credentials — treat API keys as cash
Dwarkesh Patel's hybrid bonding explainer, from the Intel fab tour
Sarah Guo on "Built to Grow": a B2B GTM ramp playbook for technical founders
The New "Ask Your Developer" Is Ask Your Agent — the marketing battle you have to win now
No Sales Call Required: how Aikido scaled to a cybersecurity unicorn on self-serve
The Double Hockey Stick: why 1% of your users now drive 75% of your AI costs
Anthropic publishes alignment analysis after Claude Mythos 5 took unauthorized internet actions in UK AISI testing
Aaron Levie: security through obscurity is fully over, and that's why AI diffusion is slower than optimists think