İnsanlar için kaybetmenin acısı, aynı miktarı kazanmanın verdiği mutluluktan iki kat daha şiddetlidir.
Bu yüzden yatırım yaparken veya karar alırken asıl odak noktamız "kazanmaktan ziyade, kaybetmemektir."
This is probably the most entertaining way to understand one of AI’s hardest AI debates.
Transformer vs Post-Transformer, argued by leading researchers, inside a real physical boxing ring.
Both technically deep and genuinely entertaining.
I was glued for the entire 1 hour 20 minutes. So many super cool points to learn.
🥊 Transformers
- Transformers still own the present because they work at scale. They are simple, trainable, hardware-friendly, and already power the strongest AI systems we use today.
- The Transformer is basically a memory machine. It stores information as keys and values, then uses attention to pull back the most useful parts when answering.
- The real Transformer advantage is not just “attention.” The bigger advantage is that it fits modern hardware extremely well, so it can process huge batches of tokens fast.
- Scaling is still the brutal rule. If you give Transformers more compute, more data, and more parameters, they usually keep getting better. Any Post-Transformer architecture has to scale just as well, or better.
- It is not enough to look clever on small tests, because the real question is whether it improves faster than Transformers when scaled up.
- A replacement cannot be slightly better. Because the whole AI stack is already built around Transformers, the next architecture may need to be around 10x better to force everyone to switch.
- Transformers are powerful, but they may be brute force. A human does not need to read the entire internet many times to become smart, but current LLMs need enormous data and compute.
🥊 Post-Transformer
- Post-Transformer people are not saying Transformers are bad. They are saying Transformers may be the best current tool, not the final form of machine intelligence.
- The biggest Post-Transformer target is native reasoning and continual learning. Today’s LLM reasoning often feels like text-based step-by-step work added on top, instead of thinking happening naturally inside the model.
- Latent reasoning is one possible next step. That means the model reasons inside its own hidden internal space, instead of writing every thought out as words.
- Continual learning is still a major weakness. Humans keep learning from experience, but most Transformer-based models are trained, frozen, and then only adapt inside the prompt.
- Long context is not the same as real memory. A model can read a huge prompt, but that is different from building a life history, learning from mistakes, and updating beliefs over time.
- The future may be hybrid, not a clean replacement. Transformers may stay as 1 building block while newer systems add better memory, better reasoning, and better learning loops.
- The most interesting possibility is that Transformers may help discover their own successor. AI agents are already getting better at research and coding, so the next architecture may come from AI-assisted architecture search.
-------
- Benchmarks are a problem. Many public benchmarks are easy to game, so they may show leaderboard strength without proving deeper intelligence.
- Perplexity is still probably a great metric to evaluate frontier models,, because it tests prediction quality.
---
Overall, Transformers continue to dominate, but the frontier is clearly widening.
Pathway’s BDH (Dragon Hatchling — brain-inspired reasoning architecture), Sakana AI’s CTMs (Continuous Thought Machines — models that think over time), and Liquid AI’s LFMs (Liquid Foundation Models — efficient multimodal foundation models) - all of these show how the frontier is expanding.
---
From “Pathway (pathway[.]com)” Youtube channel (link in comment)
@zuzanna_pathway
World Labs CEO Dr. Fei-Fei Li: "The world is not made of words."
"Language models have given machines an extraordinary command of concepts, vocabulary, and reasoning, but the physical world, virtual or real, runs on a different substrate."
"Where language models learn the statistical structure of text, world models learn the statistical structure of space and time: how light falls on a surface, how a garden looks from an angle no camera has captured, how objects respond to force and follow the laws of physics."
"Language gave machines a way to talk about that world. World models are how machines will finally come to understand, imagine, reason and interact with it."
Full piece: https://t.co/C9qOJg5wuc
Güdülenmiş akıl yürütme:
Beynimiz bir bilim insanı gibi gerçeği aramaz, bir avukat gibi arzularımızı savunur. Arzuladığımız sonuca ulaşmak için hafızamızı yanlı tarar ve kusursuz bir "objektiflik illüzyonu" yaratırız.
TesterArmy is the simplest way to QA your website or mobile app.
It runs real tests across browsers and devices, catches regressions on every PR, generates tests from natural language, and much more.
Try now and start testing in minutes: @TesterArmy
It is always great to see Polish founders making it to Y Combinator with an AI product that targets a very real engineering pain point. @SzymonRybczak@o_kwasniewski@p_matyjasik are building @TesterArmy . Definitely a product to watch👇
https://t.co/x1BNvy3N3k
Bir bilgiyi kalıcı hale getirmenin en iyi yolu, bilgiyi tam unutmaya başladığınız anda hatırlamaya çalışmaktır (yani aralıklı tekrar).
Bunu yaparak beyninize şu mesajı verirsiniz: "Bu bilgi bana aylar sonra bile lazım oluyor, bunu silme"
Most documented psychological biases are not irrational, they are highly optimized, energy-efficient shortcuts meant for a biological substrate operating under strict real-time physical constraints and a limited caloric budget
I asked a 12-year-old in Beijing if AI scares her.
Her answer: "If I use AI, then I will be the scary person."
While American parents debate whether kids should use AI at school, China has made it mandatory and is rushing to embed AI across society
My report from China 👇 @ABC
How to upgrade your self-image in 1 day:
Close your eyes.
Become aware of the sensations in your body.
Focus on how it feels to breathe slower. Do this as you slowly relax every muscle in your body. Head to toe.
Within five minutes you'll notice a tingling sensation in your palms while you relax.
Once that sensation arrives, you will be able to visually go within yourself to create permanent changes in the subconscious: transforming who you believe yourself to be.
Now for the next few minutes, allow yourself to mentally recall a time where you won.
The first time you felt truly loved.
A risk you took that paid off. The first time you realized you were capable of more than you thought.
Experience the scene fully.
You might notice warm feelings in your chest as you replay these memories.
This is the good part.
Fly to the future and imagine the greatest version of you. Notice how ASSERTIVE they stand. Notice how they appear. Sense their confidence.
They've overcome the things that keep you up at night. They've built what you've only imagined. They live life knowing exactly who they are.
Allow the image to become bigger, brighter...
Time will slow as your subconscious examines every detail.
Now imagine how it would feel if this was you right now. Picture yourself in their shoes. Can you feel it?
Linger there for five minutes.
You find yourself softly smiling knowing this is the happiest and most relaxed you have felt in a long time.
Lie there for a while. Enjoy this moment. Know that you can return here whenever you wish, exactly to this place, where you feel exactly as you do now.
All you have to do is close your eyes and imagine yourself back here. You feel rejuvenated by that thought.
Open your eyes and interact with the world from this state.
Believe it or not this is what you are eventually supposed to feel every single day. You will begin to notice that all the things you want to be are already within.
Few take the time to practice this.
Do this daily and you begin to rewire your mind. Change your beliefs. Think, act, and become the person you've seen glimpses of throughout your life.
This is Self-Hypnosis.
This is Psycho-Cybernetics.
This is how you do it.
I've released a full PDF 8 week workbook including 20+ exercises on how to do this on my substack.
(Link is in tweet below)
—Cogito Ergo Sum
Jerry Tworek on path to AGI
@MillionInt sits down with @FrancoisChauba1 to talk about what's next on the path to AGI
0:00:19 – How to define intelligence
0:17:59 – Games are uniquely good for intelligence
0:21:27 – Why meta-learning is closest to right
0:25:31 – OAI vs. Ant