A3A
25 posts
A3A
@abdabdTECH
Interested in ML SWE ------- PostDoc. Material Science (Uni. Paris Saclay) PhD. Electrical Engineering (Univ. Paris Saclay) MsC. Acoustics MsC. Mechanics
France Katılım Ağustos 2025
36 Takip Edilen2 Takipçiler
what the fuck?
i asked ChatGPT to help me optimize a programming algorithm
and it thinks i'm suicidal ?
English
GPT-5 is the smartest model currently, though others may find specific use cases
chatGPT is slightly better overall, with most competitors copying its features and UI, memory, voice, and img-gen
for coding, codex, and claude code
openAI is at the frontier of almost everything
English
English
@alex_prompter I agree, here is my experience with it. Not only does it affect intelligence, but over time, it becomes less productive to use it.
English
everyone's racing to use AI faster.
nobody's asking what it's doing to their brain.
i just read a 132-page research paper that should terrify every creator, marketer, and founder using AI right now.
it's called "The Impact of Artificial Intelligence on Human Thought" and it explains why most people using AI are accidentally making themselves dumber.
here's the problem: when you outsource thinking to AI, your brain stops doing the work.
the researchers call it "cognitive offloading" - basically, mental atrophy.
you think you're being efficient. you're actually losing the skill that made you valuable.
the worst part? it's invisible. you don't notice your critical thinking weakening until you try to solve something without AI and... can't.
here's what the research actually says:
English
we finally know how close we are to AGI
the paper tested the GPT-4 and GPT-5 on human cognitive abilities:
> GPT-5 scored 58% toward AGI
> GPT-4 scored 27%
the research also shows "jagged intelligence", which helps explain why AI can feel both very impressive and surprisingly weak at the same time
English
I don’t think ANYONE is ready for when AI starts predicting our next thoughts and actions…🔮😏
English
@elder_plinius AI is deterministic like it or not, human behavior on the hand is.... at least doesn't seem deterministic.
English
Larry Ellison says today's AI models are no longer just inferencing; they reason like we do
These systems handle deduction, calculation, and strategy far faster than people
"at Oracle, most code is now AI-generated"
We declare our intent; the model writes the procedure
English
@davidasinclair Living long lives, is it because life is good or are we just afraid of death.
English
The first generation of humans to control their biological age is alive today. Think about that
English
@Yesterday_work_ Unless the weights can change after training, it won't make a revolution.
English
🚨 MIT just built an AI that rewrites its own code to evolve
It’s called SEAL (Self-Adapting Language Model) and it might be the first real step toward self-improving intelligence.
Instead of waiting for humans to fine-tune it, SEAL:
- reads new information
- rewrites it in its own words
- performs gradient updates *on itself*
Literally… it learns how to learn.
The results are insane:
✅ +40% boost in factual recall
✅ Outperforms GPT-4.1 using data it generated itself
✅ Masters new tasks without human input
LLMs that train themselves aren’t sci-fi anymore.
We just entered the era of self-evolving AI.
Paper: jyopari. github. io/posts/seal
English
@koltregaskes Ai like all other 'finite systems' will eventually saturate. AGI can't be achieved with deterministic models.
But maybe allowing the weights to change after learning would be a good start.
English
This was classed as a level 4 achievement by OpenAI... out of 5, before we got "AGI".
The next couple of years could be incredible.
A lot of the new AI data centers will come online in 2027-2028+. Power will be one bottleneck then human resistance.
But (my personal bugbear) floor capabilities are improving now; hallucinations, accuracy, etc. Scaling still shows progress, models get cheaper (but used more!).
I'm not saying the current Transformer/LLM architecture is it, perhaps there is something new needed, and we do have some biggies outstanding, such as continuous learning, infinite memory etc.
But beyond that, nothing is stopping superintelligence arriving.
IMO. 👍
Sundar Pichai@sundarpichai
An exciting milestone for AI in science: Our C2S-Scale 27B foundation model, built with @Yale and based on Gemma, generated a novel hypothesis about cancer cellular behavior, which scientists experimentally validated in living cells. With more preclinical and clinical tests, this discovery may reveal a promising new pathway for developing therapies to fight cancer.
English
Be honest!
Does this setup look like $250,000 Bitcoin $BTC by December?
English
$BTC
Monday High ✔️
Tuesday Low ✔️
Monday High was textbook. Tuesday Low formed through NY open sweep rather than Asia! This is why observing sessions is important to time the pivots!
x.com/KillaXBT/statu…
Killa@KillaXBT
$BTC Remember my pivots! Monday High. Observe during London/NY for the high to be formed. In some rare cases, the high is formed during Asia. But most cases during London & NY.
English
$BTC
As we can see, on this recent drop, orderbook delta has flipped positive. A sign that buyer are accumulating.
Since september, every time delta has flipped green we have seen a push upwards on the HTF. Will this time be any different?
x.com/KillaXBT/statu…
Killa@KillaXBT
$BTC As I said, still a taker dominant market. Nobody was bidding. It was shorts covering on the move to 112.4K. Simply put, they took out late shorts, built late long liquidity & wiped out both directions. x.com/KillaXBT/statu…
English
@the_yanco I don't think AGI will ever come. AI is really not that 'smart'. it is all about recognizing patterns.
Which btw classical physics can already do a billion times better than AI.
And humans can do much much more than simply recognizing patterns.
English
If LLMs in existed 1902 everybody be like:
"We cannot have heavier than air flight without AGI!"
Carl Hasselskog@CarlHasselskog
@the_yanco @bryan_johnson Aging is already killing everyone.
English
@Dr_Singularity Remember AI goal is to minimize a loss function and if by offing all humans the cancer loss function is reduced AI might just converge to that solution.
English
@AskPerplexity Let it happen, everyone needs to see how dumb AI is, and that we are no way reaching AGI of cat let alone a human.
English
Oxford researchers just confirmed what we feared:
The internet as we knew it is dying.
AI content went from ~5% in 2020 to 48% by May 2025. Projections say 90%+ by next year.
Why? AI articles cost <$0.01. Human writers cost $10-100.
But the real crisis is model collapse. When AI trains on AI-generated content, quality degrades like photocopying a photocopy. Rare ideas disappear. Everything converges to generic sameness.
It's recursive. Today's AI slop becomes tomorrow's training data, producing worse output, which becomes training data again.
English
