AI tech briefing: Anthropic alarm, GitHub outages, Skild robots learn from video

Jacob Coxin quit Anthropic at 27 warning the builders fear extinction by 2030. This AI tech briefing also covers GitHub's rough August and Skild AI's video learning robots with NVIDIA.

AI could kill all humans by 2030

Coxin said the people building AI earnestly believe it could kill everyone by decade's end.

He framed it as private consensus, not marketing. Evan Hubinger, an Anthropic alignment scientist, shared the post and endorsed the warning.

Source: YouTube

Skild AI teaches robots new tasks from a single video

S1 is a robot foundation model that executes long horizon tasks from one video demonstration.

It uses in context learning with no weight updates or task specific post training.

Skild built and trained S1 on NVIDIA AI infrastructure spanning synthetic data generation and simulation.

That matters because factories change layouts and products constantly. Video in removes reprogramming cost for each new task.

Source: NVIDIA Blog

GitHub said August 2026 was challenging for availability despite architectural work and a move to Azure.

Growth continues and risk cannot be eliminated. Repair items focused on capacity improvements.

Source: Github Blog

StartupHub data

Anthropic

Anthropic

Anthropic is an AI safety and research company building reliable, interpretable, and steerable AI systems, best known for the Claude family of models.

Founded

2021

Location

San Francisco, California, USA

Valuation

Private / $100B+ est