always curious about scalability, big numbers, golden ratio; attracted by beauty, fine observer of Peter's principle
Helsinki, Finland š«š® š·š“
Joined November 2013
- Tweets9.9K
- Following360
- Followers173
- Likes2.8K
Yo world. Want to hear how Apple Intelligence (on device) iSee you? apps.apple.com/fi/app/isee-yā¦
Ciprian Manea retweeted
šØ: Quantum entanglement visualised for the first time ever
Readers added context they thought people might want to know
Firstly, this is not new news, its from 2023.
Secondly, the yin-yang pattern isn't the shape of atoms or atoms entangling. Its simply a picture that was used by the team as a test object for this holographic imaging technique. Any image could have been used. Important context..
livescience.com/physics-mathemā¦
Ciprian Manea retweeted
Unrelated, I'm sure š
NVIDIA āunlaunchesā its GeForce 595.59 driver after reports of fan and clock issues
videocardz.com/newz/nvidia-uā¦
Ciprian Manea retweeted
"Learning is rooted in repetition and convexity, meaning that the reading of a single text twice is more profitable than reading two different things once." - Nassim Nicholas Taleb in Skin in the Game
Ciprian Manea retweeted
What all happens when you verify your LinkedIn identity?
thelocalstack.eu/posts/linkeā¦
Ciprian Manea retweeted
Nearly every board meeting : "Hiring strong infra folks is incredibly hard right now"
Ciprian Manea retweeted
Even worse - each neuron that contributes to your cognition and consciousness might contain 100,000 Microtubules - each Microtubule Operating at 10,000,000 oscillations per second - and being Quantum photonic memristers
Good luck with the AGI tho
Ciprian Manea retweeted
AGI is not coming.
We are nowhere near AGI. What we have today is inference, not learning.
Models get trained once on huge fixed datasets, then frozen. You ask questions, they remix patterns they already saw. Nothing updates. Nothing sticks. Talking to the model does not make it smarter. It does not learn from you. Ever.
Learning is still slow, expensive - and offline.
Look at self driving. You drive around a pothole, make a U turn, and come back. The carās AI does not learn that you just solved that exact problem. It reacts the same way every time using sensors and rules. Do this 20 times a day and it still has zero memory that the pothole exists. It just re sees it. That is why edge cases never die. There is no local learning. No accumulation.
No 'oh yeah, Iāve seen this before'
LLMs work the same way. Tell it your name and it does not remember. The only reason it looks like memory is because scaffolding keeps shoving your name back into the prompt every time and sanitizing the output.
The model itself has no idea who you are and cannot learn from interaction. It is structurally incapable.
And the scaffolding is the worst part. It is pure duct tape. Just prompts on prompts on prompts around a frozen model. When something breaks, nobody fixes learning. They add another layer. Another rule. Another retry. Another evaluator model judging the first model.
So you end up with systems that are insanely complex but mentally shallow. Debugging is hell because behavior comes from hack interactions, not a learnable core. Tiny prompt tweaks cause wild behavior shifts. Latency goes up. Costs go up. Reliability goes down. None of this compounds into intelligence. It just hides the cracks.
Until we have real persistent learning and real memory inside the system, there is no AGI.
LLMs are not built for this. You cannot prompt your way out of it. You need a totally different architecture. Yann LeCun is right.
And even then, what architecture can actually learn online, store memory, and stay stable on todayās hardware?
Best case, maybe 5-10 yrs.
Right now it is all inference. It looks magical, but the emperor has no clothes. A lot of people see it. Almost nobody says it out loud.
Ciprian Manea retweeted
If there was a school of Silicon Valley, this would be required reading.
@WillManidis: thank you for writing this.
Ciprian Manea retweeted
Kurt Metzger brought up a point to me today, he said the Epstein Files are a test of compliancy, if there is a bit of outrage but overall it simmers down and is accepted without major consequences, if humanity allows this, then permission for the next phase is given.
Ciprian Manea retweeted
The fastest way to validate a product:
DON'T build it
-Make a landing page
-Describe the problem it solves
-Add a "join waitlist" button
If nobody signs up, you just saved yourself 3 months
Ciprian Manea retweeted
Weāre industrializing how software is made. I use āindustrializeā deliberately. When you industrialize a product, you flatten it so it can scale. You want a Heineken to taste the same everywhere in the world, so you standardize. You lower the bar so everyone can hit it. Thatās whatās happening with AI.
Iām not judging this as good or bad. It just is. It happened with food, with fashion, with drinks. Now itās happening with software. Weāre entering an era of industrialized software production. Weāll have far more software than ever, but a lot of it will be average.
Ciprian Manea retweeted
Updated Post: Apple: iPhone Demand Stole The Show, But AI Distribution Is The Catalyst asymco.com/2026/02/03/apple-ā¦
