Human brain and all the complexities that the brain seamlessly elucidate is one of the most fascinating thing for me to understand.
So here is a 🧵 of articles, talks and videos pertaining to this topic that I found riveting.
[1/n]
@AirEuropa - I am trying to obtain an invoice for one of my previous travels but I keep getting some error asking me to contact the customer service directly. I remember accessing the invoice without any problems previously for the same reference number but it is possible now!
Last Sunday at Gorakhpur airport.
My sister had a spelling mistake on her boarding pass. Just the last letter of her name.
Security stopped her. Sent us to the counter. Counter said call Akasa support.
We called. AI agent picks up.
We said the PNR like 5–6 times. AI couldn’t recognise it. Not even once.
We’re standing at the airport, stressed, flight’s approaching and this bot is just not getting it.
We gave up. Went back. Shouted at the staff. They fixed it in 30–40 mins manually.
And I just stood there thinking.
Bro who thought deploying an AI agent for customer support was a good idea?
The person calling support is already having a bad day. They’re not calling to chat. They’re calling because something went wrong.
That’s the worst time to make them repeat a PNR to a bot 6 times.
Some problems just need a human on the other side.
AI is powerful. But knowing where not to use it is the real skill.
Founders take note.
@anilkseth With the criticism revolving around the fact that there is no evidence for how such inferences arise in neuronal-level. I'm curious to go through any papers or literature that try to explain this. @anilkseth please help me with this, and would also like to hear your thoughts too.
A few months ago, I had detailed out my understanding of @anilkseth's "Being You". Going through this essay on conscious AI reminded me once again of our incomplete understanding of what it means to be "us" and our tendency to associate consciousness with intelligence. (1/n)
1/ I am truly honoured and completely delighted to have won the 2025 Berggruen Essay Prize (English), for my submission “The Mythology of Conscious AI” @berggruenInst@NoemaMag https://t.co/n3eXSRXkvk
Recently, I've read literature on predictive processing and Bayesian inference (especially in language processing). My biggest question is how Bayesian inference arises in biological substrates. I also came across some papers that criticize this view. @anilkseth (2/n)
🎊📜 NEW PAPER 📜🎊
Can we seriously build synthetic consciousness?
And if so, where do we start?
I’m super excited to present recent publication in Neuroscience & Biobehavioral Reviews where @jaaanaru and I confront this challenge head on.
https://t.co/6Jk9Ce6eE5
1/n
Ended up reading a thread full of anti- and pro-AI views (or maybe more like anti- and pro-LLM). Probably the only serious thought I have these days is if EVERYTHING is automated, then what would humans even do on day-to-day basis? Just eat and sleep?
I am not against AI per se. It is arguably the revolution of our time as it was the Internet before and electricity before that.
However, we are all increasingly seeing an indiscriminate use of AI as a way of advertising already existing products with the little extra feature that it can tell you "Good morning, do you want your espresso as always, Stephanie?". Companies use a make up mask over their stock in an intent of making the customers believe they have made a significant upgrade in their product.
A perfect example of a brand that has already made this move is Apple with its AI. Maybe I am mistaken, but I would love to hear someone explaining to me a useful feature of AppleAI that isn't making an emoji with the face of your brother. And then, I would love seeing some statistics about how much do customers use AppleAI compared with the huge amounts of advertising Apple made of its old products with the only new feature of AI. It is not an isolated example: you can see loads of minor software brands trying to make you pay a monthly subscription so you can use an internal AI on their product. As curious as it may seem, most of these brands are desperately seeking you to buy a subscription of whatever feature they offer, and AI is the perfect opportunity for them to chain you with $9,99 a month.
You didn't put electricity on a baseball bat in the 30s but you are adding AI on your fridge today.
For sure AI is useful on stuff. For example in vehicles, especially on those with computer vision integrated as Tesla's autonomous driving vehicles. But I am trying so hard to imagine Elon Musk not putting a feature on the gigantic screen where you can ask Grok if Joe Biden would win in a fight against 3 furious puppies.
AI could be used in so many creative ways that would actually enhance our interaction with technology and make us step forward exponentially quicker. Both for tech experts and for the general public all over the world. But today's productive system is tricky. The objective is not to advance forward, it is that you buy whatever makes you think you are doing so. Sure you can understand why this happens. But understanding it doesn't make it better.
You are going to end up with an emoji of your mom with butterfly wings when you could have a phone that could guide you step by step on how to build your own SLM. Sure you can already get both today if you look hard enough in small places on the internet. But which one are companies promoting? They need those $9,99 a month for a feature that is mostly worthless.
I am tired of seeing mug warmers that require AI to find your preferred coffee temperature and companies with dollar signs in their eyes making you think you need those mug warmers and milking you.
Only two English sides have won the UEFA Cup/Europa League three times:
◎ Liverpool
🏆1973
🏆1976
🏆2001
◉ Tottenham Hotspur
🏆1972
🏆1984
🏆2025
Sevilla (7) are the only side from any country to win the competition on more occasions.
This is still one of my favorite #dataviz of all time.
It shows how sensitive results are in small samples.
It provides raincloud plots with N=20 (on the left) & N=500 (right).
There is no difference between the two conditions other than the N.
Before we start discussing #FreedomOfSpeech we should agree that Telegram is the Dark Web 3.0 and their platform enables an infinite number of all kind of criminals.
It is Not a freedom of speech when an anonymous username sells drugs or stolen credit cards 💳 leveraging security features of the Telegram platform.
It is not a freedom of speech when known jihadists leverage the platform to plan another attack on church or Christmas market.
It is a matter of public safety when LEAs have access to data that helps them to prevent such attacks and crime.
Freedom of speech is when you have a first name and last name, and you free to post your opinion or become a citizen journalist to share updates/news/concerns with your followers without the fear of prosecution.