Holy crap! Anthropic IPO Prospectus Reveals $8 Billion Operating Loss. You'd think the company was swimming in cash, but no. Think it'll matter, though?
––– 良くない–––
Misleading AI-generated summaries can distort human memory. Turns out that when humans are presented with faulty AI summaries of meetings and such, they're more willing to say "Uh, OK..." rather than "Wait, that didn't happen!" This is scary, especially when AI is used to summarize clinic encounters. How often does your doctor correct errors, or just assume it happened that way? What about when Apple watches summarize recorded events. Humans have never had to face this before.
––– 良くない–––
Researchers at Google DeepMind have found a way to watermark AI-generated proteins without affecting their function. This is a way to track the provenance of their proteins. Paper here. Reminds me of Blade Runner 2029, where you can read the serial number of replicants.
––– 凄い –––
Chinese publishers announced the creation of their new journal, Vita, which they aim to make as prestigious as western journals. To encourage westerners to submit, they are waiving the article publishing costs, for now. We'll see how this goes.
––– 凄い –––
The rumors are true: Reddit is sharply curtailing access to old.reddit.com and is killing RSS feed access as of November 13. All because of AI scraping and content bots. This will probably kill the allure of Reddit, and once again encourage alternate social media sites. We'll see. This would be an opportunity for Digg.com to come back, after their unsuccessful attempt at a comeback.
––– 良くない–––
OpenAI holds off on releasing new model over safety concerns, saying it "didn't quite meet the bar". I think frontier models are getting so advanced now that their should eliminate unrestricted release to the public.
––– 凄い –––
September 2026: the world today, as seen by one Polish guy. Excellent summary of the energy crisis that Poles face, and how it's interconnected with world politics. The country is in a tight spot, and they're relying on a lot of luck, since they haven't prepared for disaster, and are suffering from the decisions of the EU, of which they are a member. The visuals are superb.
––– 凄い –––
Oops. AI models keep posting screenshots showing sensitive data from inside tech companies. No wonder when their employees use frontier models to ask questions. I've wondered why stuff like this wasn't uncovered before.
––– 良くない–––
New Zealand teachers union demands that teachers not be replaced by AI. Coupled with my post yesterday, I would say they're arguing from a weak stance. Portland's teachers should take note.
––– 良くない–––
The latest release of Firefox has changed the browser's appearance. The working viewspace is more spacious and the corners are rounded. I've had to modify my userChrome.css to make things they way I like them. But so far, no major change in functionality.
––– 凄い –––
Seattle is set to adopt eye-popping minimum wage amid business exodus. Not to mention the millionaire tax.
––– 良くない–––