@discord_support did write you a DM three days ago, but got no response. Would need some help, since my account got deactivated directly after my first login 🙄🥱
“The Hormuz Strait will have to be guarded and policed… by other Nations — The United States does not!”
If taken at face value, this is how you get Black Monday-type pricing in markets. But the weekend gives plenty of time for him to reconsider.
Great work - but now imagine what we can achieve in the near future with specific trainee AI models, like Google DeepMind.
Right now, psychiatry is developed like medicine in the middle ages. Just like doctors back then could diagnose "fever" but had no idea about viruses, bacteria, etc., we as psychiatrists are diagnosing "depression" without knowing the root cause.We do know that there is more then a simple Deficiencies of neurotransmitters - immun diseases, gut microbiome, genetic causes, and many others - but in reality we know too little to really integrate those diagnostic ways to apply individually tailored therapy..
My hope is that science, with the help of AI, will solve all these mysteries, and we will finally get to psychiatry, where other fields are, and we will not simply diagnose "depression" but will be able to diagnose it through new light, more accurately understand the causes, and be able to treat them.
🧬 What is the link between genetics and which part of the brain is affected by mental illness?
Read how our researchers are investigating this question in our new blog for #BrainAwarenessWeek.
https://t.co/RAC7h9aNkK
@NIHRMaudsleyBRC
🧬 What is the link between genetics and which part of the brain is affected by mental illness?
Read how our researchers are investigating this question in our new blog for #BrainAwarenessWeek.
https://t.co/RAC7h9aNkK
@NIHRMaudsleyBRC
Huawei and Oracle are both framing agentic AI as the next infrastructure layer at MWC — autonomous systems orchestrating networks, finance ops, and indoor intelligence.
The excitement is understandable: faster decisions, seamless execution, less human friction.
What gets overlooked is the psychiatric downstream: when agents generate recommendations, content, or even "insights" independently, the origin of every output becomes one more thing to verify.
We are not just accelerating workflows. We are multiplying the occasions where source monitoring must intervene — in real time, indefinitely.
Velocity wins the day. Veracity pays the compounding interest. And our epistemic immune systems were never built for perpetual audit.
In recent days, the developments in Iran have coincided with a marked rise in AI-generated visual material across platforms.
Images purporting to show specific scenes, fabricated satellite imagery, and synthetic footage of military activity have circulated widely, frequently before independent checks could be completed.
As a psychiatrist, what I notice is the steady, incremental demand this places on source monitoring. Our cognitive systems are now routinely asked to separate genuine from generated input under conditions they were not shaped for. This is constantly puting more stress on our cognitive systems.
Velocity has again moved ahead of veracity. Markets are pricing the immediate effects on oil and regional stability. The subtler, ongoing shift in how populations establish basic trust in visual evidence remains, for now, largely unpriced.
Not a crisis, simply a quiet recalibration of our shared epistemic environment.
You don't need to be fooled by a deepfake for deepfakes to damage you.
Researchers call it the "liar's dividend": once you know synthetic media exists, you start doubting authentic media too. Not because you've been deceived — but because you could be. The mere possibility is enough.
This is the part most AI commentary misses. The threat isn't just that bad actors will create convincing fakes. It's that the existence of convincing fakes degrades your trust in everything — including things that are real.
Think about it: when's the last time you saw a video online and your first reaction was "is this real?" Five years ago, that thought didn't cross your mind. Now it's automatic. That shift didn't happen because you were tricked. It happened because your brain updated its priors.
As a psychiatrist, this interests me because the phenomenology is eerily close to derealisation — the clinical experience that your surroundings feel unreal, distorted, dreamlike. Classic derealisation is a perceptual distortion the patient recognises as abnormal. But AI-induced reality distrust is different. It's worse, actually. Because the uncertainty is objectively justified. You genuinely can't tell. That's not a malfunction of perception — it's an accurate read of an increasingly unreliable information environment.
We're building a world where rational people experience reality the way derealized patients do — and we don't have a name for it yet.
Your brain doesn't have an AI detector.
The fusiform face area — the neural circuit that processes faces — treats a deepfake exactly like a real person. Same activation, same automatic trust response. Your gut says "real" before your cortex even gets a vote.
Overriding that takes conscious effort. Every time. For every face, every voice, every piece of text you encounter. Cognitive scientists call this source monitoring — the process of figuring out where information came from. It's handled by your prefrontal cortex, and it's effortful, slow, and capacity-limited.
Now imagine running that process continuously, for every piece of content you consume, indefinitely.
That's not a metaphor for cognitive load. That IS cognitive load — of a kind we've never had to carry before. A chronic tax on the exact brain systems that keep you grounded in reality.
When source monitoring fails in clinical populations, we call it psychosis. When it's overwhelmed in healthy populations by an environment that demands it nonstop — we don't have a name for that yet.
But maybe we will.
Everyone's debating whether AI will take your job.
The bigger threat? It's taking your sanity.
I'm a psychiatrist, and here's what I think the tech discourse is missing entirely:
Every major technological revolution has produced psychiatric casualties. Industrialisation brought mass alienation and alcoholism. Social media gave us an adolescent mental health crisis we're still not done processing. In every case, medicine was reactive — we documented the damage after the fact.
AI is different. And not for the reasons people think.
The internet increased the velocity of information. AI undermines the veracity of information. That's not a quantitative upgrade — it's a qualitative shift that hits the deepest layers of human cognition.
Our brains run on what you might call an epistemic immune system — the ability to sort trustworthy from untrustworthy input. It's not a luxury. It's a core function. When you can no longer tell what's real — when any image, voice, video, or text could be synthetic — that system starts to fail.
This isn't information overload. It's epistemic destabilisation.
The downstream effects are predictable: chronic uncertainty erodes reality testing. Allostatic load rises. Anxiety, dissociation, paranoid ideation — these aren't speculative outcomes, they're the expected neuropsychiatric response to a world where the ground truth keeps shifting.
And we're not ready. Not even close.
The psychiatric profession can wait and document the casualties — as we always do — or we can act with foresight for once.
The epistemic cost of AI is unhedged. And nobody noticed.
I'm Boris and I created Claude Code. Lots of people have asked how I use Claude Code, so I wanted to show off my setup a bit.
My setup might be surprisingly vanilla! Claude Code works great out of the box, so I personally don't customize it much. There is no one correct way to use Claude Code: we intentionally build it in a way that you can use it, customize it, and hack it however you like. Each person on the Claude Code team uses it very differently.
So, here goes.
Few days remaining before we close this final Kittysquiddy accession window.
Make sure you send your submission THIS week.
End if week we will stop accepting requests to focus on processing everything.
[email protected]
@CL_Technology hey there! Got a question regarding crypto draft credit line - is it borrowed as euros or as a stablecoin? Makes a big difference, since paying with stablecoins includes selling them first to euro, which would make it a taxable event. Or not?
@kittysquiddy Attention fellow feline lovers - people keep trying to get into the group. Had follow request not only on this account (atm following @kittysquiddy ) but also on my new locked handle. Remember the Membercheck website https://t.co/Cqlv18c9zm
@kittysquiddy Done ✅ As I didn’t know how to handle the transition to the new locked group I followed from this account and also my new locked account @marandoKitty