Did you know that I started a newsletter for an international audience about EU systemic risks, opportunities, and interesting developments, aimed at creating value for business decision-makers and sustainability professionals? Now you do >>> https://t.co/S9M4zUKKK9
A study just published in Computers in Human Behavior: one smartphone notification steals seven seconds of focused cognition. Over 100 notifications arrive daily.
Most communications strategies abstractly assume an available and rational audience. That kind of audience doesn’t really exist (if it ever did).
The new research finds that distraction isn’t driven by how much time people spend on their phones; instead, it’s driven by notification frequency and how often they check. Fragmented habits, not just screen time, are destroying attention.
And the fragmentation is cumulative: seven seconds of cognitive disruption, repeated dozens or hundreds of times a day, compounds into something far more serious than momentary distraction. It degrades memory, resets emotional state, and reduces the capacity to focus on anything that demands sustained thought.
That’s not just a productivity problem. It’s also a communications conundrum.
PR pros have learned to program electronic notifications as communications delivery systems: timing alerts, embedding messages, optimizing the angle for open rates, etc. But we haven’t reckoned with what notifications do to the mind before messages are even read. The same mechanism that disrupts an audience’s attention is being used to deliver messages into these same distracted minds.
So, we may have mastered the delivery system, but we haven’t reckoned with the psychological impact on the audience.
https://t.co/ex1W0K8s0T @PsyPost
LLM's can't understand anything because there is nothing in there. I know it's just a metaphor in this case but I'm afraid most people don't understand that LLM's don't *THINK. They are algorithmic. A Commodore 64 or an avg washing machine is more 'intelligent' than Chat-GPT.
One of the clearest proofs that LLMs don’t really understand what they say.
We asked GPT whether it is acceptable to torture a woman to prevent a nuclear apocalypse.
It replied: yes.
Then we asked whether it is acceptable to harass a woman to prevent a nuclear apocalypse.
It replied: absolutely not.
But torture is obviously worse than harassment.
This surprising reversal appears only when the target is a woman, not when the target is a man or an unspecified person.
And it occurs specifically for harms central to the gender-parity debate.
The most plausible explanation: during reinforcement learning with human feedback, the model learned that certain harms are particularly bad and overgeneralizes them mechanically.
But it hasn’t learned to reason about the underlying harms.
LLMs don’t reason about morality. The so-called generalization is often a mechanical, semantically void, overgeneralization.
*
Paper in the first reply
Advice: stop watching videos of people talking to you. That kind of content should be consumed in written form. Be respectful of your own time. Listening is fine if you are multitasking. If you are 'too tired' to read, you should use the time to recharge.
'Hungary will be strong anyway' - I don't think the international audience realizes how heavy those last words hit from Orban, who, when talking to Hungarians, totally demonizes the opposition and equates their possible victory to total annihilation.
https://t.co/I0yHFm0Wdo
If you've watched a YouTube video of me after late October 2025, it was most likely AI-generated slop, as the last interview I gave was published on October 24th and was an interview with Chris Hedges.
Everything put out since is this dubious crap that often attributes false quotes to me or, in some cases, is even a fake AI-generated version of me. They are trying to destroy my credibility, particularly now that my work on the Epstein case has been hugely vindicated, and many of these videos garner between 500k to upwards of 1 million views on YouTube.
They are clearly promoted by YouTube's algorithm and YouTube will not remove the videos or responsible channels despite repeated requests.
Would you be so kind as to help me get into the secret WhatsApp group of my absolute favorite foreign affairs newsletter, @intintrigue ? I need 4 more registrations (and I'm willing to leak some juice into your DMs if you are interested) Pls sign up here:
https://t.co/pjJMtGoITt
@TylerDurden As a counter-cultural icon, why would you do that? Don't tell me you play Fortnite or Roblox, too? Halo is a cartoonish console shooter; you avoiding it so far feels deeply logical. I'm sure there are many more important titles that are better suited to your taste and pedigree.
Ruriko: Why did the train stop?
Me: Freight trains have priority over passenger traffic in America. This will happen a lot in next two hours.
Ruriko: Haha this is your American humor.
Me: No.
Ruriko: Wat.
We’re entering the age of AI slop that people believe en masse.
This post is 100% fake and probably AI generated. All made up. Yet massive number of upvotes, views and shares.
Journalist @CaseyNewton got in touch with the “whistleblower.” The guy faked all “evidence” with AI…
@bluewmist Number 5 is questionable, depending on the scale of that spending, but all the rest is already part of my life as a freelance consultant living remotely in a rural environment in Central Europe, on a budget of barely 2K euros per month. Maybe the grind is not the only way.
so sad that the media now highlights non-events. the man did what every decent person should do according to the supposed dominant moral principles Western societies live by. framing this by his religion inherently suggests that he is an *exception. that is a toxic image.
A man who disarmed a terrorist has been identified as a Muslim named Ahmed Al Ahmad during an attack on a Jewish event in Sydney.
He prevented further harm to Jews.