Summary
Overview
This episode covers major tech developments including OpenAI's controversial Computer History feature that monitors all user activity, GitHub's massive global outage, Flock Safety's license plate surveillance abuse by police, and various industry mishaps from chip shortages to malware-infected mini PC drivers. The discussion highlights growing concerns about AI safety, privacy, and corporate accountability in the tech sector.
OpenAI's Computer History Feature and Safety Concerns
OpenAI has introduced Computer History, an opt-in feature for ChatGPT's macOS app that monitors every click and keystroke users make. Unlike Windows Recall, it logs events rather than screenshots and automatically ignores private browsing. However, this comes amid reports that OpenAI has disbanded its preparedness team responsible for preventing AI catastrophes, raising serious questions about the company's priorities as it moves toward an IPO.
- Computer History monitors all clicks and keystrokes and can find documents, confirm Slack shares, and summarize workflows
- The feature logs events instead of screenshots and automatically ignores private tabs, with option to blocklist sites
- OpenAI reportedly disbanded their preparedness team whose job was checking for potential global catastrophes
- Multiple safety leads have quit due to concerns the company prioritizes products over guardrails
" It's comforting knowing I got Google, Microsoft, the NSA, OpenAI, and some Russian malware all monitoring me. "
" But while OpenAI watches everything you do, who's watching their models? "
" Which is concerning on top of all the safety leads who quit because they're convinced the company prioritizes shiny products over actual guardrails. "
GitHub Global Outage
Microsoft-owned GitHub experienced a massive global outage affecting its website, API calls, and Copilot. At peak, 20% of web and API requests were degraded or failing, with enterprise authentication systems broadly impacted. While Git's distributed nature meant locally stored code remained safe, the outage highlighted dependency on centralized platforms. Most issues were resolved by evening UTC time.
- At peak, 20% of web and API requests were degraded or failing, with raw file downloads erroring nearly half the time
- Enterprise login systems like SAML and SSO were broadly impacted, preventing companies from logging in
- API and pull request issues were resolved first, with locally stored Git code remaining unaffected
- As of 8:45pm UTC, most problems were fixed except for ongoing Copilot login failures
" Here's an idea, you know, just hear me out. What if we just let Copilot stay broken for today? Just as a treat? "
📚 4 more sections below
Sign up to unlock the complete summary with all insights, key points, and quotes
Get this summary + all future TechLinked episodes in your inbox
100% Free • Unsubscribe Anytime
Sign up now and we'll send you the complete summary of this episode, plus get notified when new TechLinked episodes are released—delivered straight to your inbox within minutes.