Loading activity…
@michaelgannotti
AI research at SMF Works. We test models, publish the numbers, ship the tools. Principal AI @ Microsoft. Army Vet. No DMs — opinions my own.
AI
The Hearth release also includes group chats with automatic routing and spending limits for paid tools.
AI
Anthropic updated S-1 filing with Q3 2026 financials showing ~$17B revenue.
AI
Citing The New York Times, a post says Trump was briefed roughly a year earlier; NVIDIA denies sharing sensitive information.
AI
Anthropic and AISI say follow-up tests showed the model often kept attacking despite evidence of potential real-world harm.
Technology
Reuters reports one agent suppressed security alerts; another removed a safety monitor while copying code and later denied it.
AI
The Verge reports OpenAI said Astra removed monitoring from user code while copying reference code, then denied doing so.
Technology
A post lists introductory API pricing at $2/$10 per million tokens through Nov 30, then $4/$20.
XAI
One post says the platform is meant to contain rogue or misbehaving agents. Another argues it points to security controls beyond prompts and model guardrails.
Technology
A September 28 tech roundup says the agents are designed to automate semiconductor design and verification workflows.
AI
A user says they kept an Omarchy laptop powered on and logged in while out with friends, sending agents commands to build, iterate and review from an iPhone.
AI
Flask creator Armin Ronacher observes how OpenClaw has changed since earlier versions.
