12 October 2026

https://logos.co/media/article/browser-betrayed-you
Are Large Language Models a Cognitive Virus?

Therefore, a gradual increase in LLM adoption can produce a disproportionate collective response. Beyond a critical point, the loss of autonomy becomes self-reinforcing and the population can move rapidly toward a state of much stronger cognitive offloading and lower cognitive competence. The important result is not that such a runaway must occur, but that it can occur under plausible forms of social learning and cooperative reinforcement. Moreover, once such a state has become established, simply returning conditions to where they were before the transition may not be sufficient to restore the previous state. Prevention can therefore be considerably easier than reversal.

Once that happens, there's no going back on AI. And if that happens, there's no question we'll need datacenters. 
And here's a nice piece showing that indeed, datacenter technology is evolving to address the public's objections.

––– 凄い –––

Two articles on browser fingerprinting:
I'm doing better than most people, it says, but it's not enough. Even the Brave and Mullvad Browsers don't fully protect against fingerprinting. The only real defense is different computers and different browsers for different tasks. Anonymous crypto transactions should be done on a dedicated computer. Same with any other sensitive browsing. Anything where you log in with credentials – same thing.

––– 良くない–––

Leading Chinese mathematicians unite behind use of AI in research. The Chinese are not afraid like American mathematicians. Probably because they don't have at much at stake. Here's a post about what the latest OpenAI dump did. Yeah, it's clear that human mathematicians will still have stuff to stay busy with. The verification process isn't as sexy but it still needs to be done.
––– 凄い –––

Hackers heat French supermarket fridges, sparking fires and spoiling food. Network security is not optional.

––– 良くない–––

Heh. Here's a holiday excuse calendar for Chinese factory suppliers – why that shipment will be late. It's always some holiday, isn't it?

––– 凄い –––

Yet another all-local (never leaves your machine) voice-to-text app. This one is BetterWispr, and it doesn't use OpenAI's weights by default, but rather Parakeet. But you can use Whisper if you want. I've decided to use it, and so far it's working great.

––– 凄い –––

Could very severe long COVID be linked to differences in lung and mouth microbiomes. "Among the changes associated with long COVID symptoms was an overabundance of bacteria that produce lactic acid. " Lactic acid seems like too simplistic a mechanism. 

––– 良くない–––

Can Wearables Really Track Blood Pressure Without the Cuff? The conclusions of the article are disappointingly vague, but it seems that they won't replace the cuff. They just provide estimates.

––– 良くない–––

ask.sciencemadereadable.com is an effort to answer science questions in a readable simpler format. In reality, the output is hard to read. At least they could have used Markdown or some HTML formatting. I'd rather use Claude or Astra to answer a science question, or even an open-source model. It would be easier to read than this site. But it's free, so that's the advantage.

––– 凄い –––
Fewer workers commute from out of state, another sign of Oregon’s soft labor market. Even SW Washington folks don't want to work in Portland. Also from the Oregonian: Gen Z sees ‘more opportunity’ outside Oregon. Hard to get a good job when businesses are fleeing. But another reason why Oregonians can't get a job is that Oregon had the highest rate of mental illness in 2023–2024. Yeah, I look around me and I can see it. Some even hold political office.
––– 良くない–––