@oelayat yep, for some people solving some] hard problems == AGI. Hassabis knows that AGI is far more than that of course.
as long as we don't have a commonly accepted definition of AGI, opinions like these will be increasing
amazonia's work is amazing.
but to push back a bit: what's stopping anyone from making stuff like this is much more than better tools.
better tools are the catalyst.
this level of work is the result of talent, dedication, and craftsmanship that no tool can ever replace.
Good post describing the toxicity of all social media sites that chose to monetize attention. What Brett describes applies to any topic, not just design-related posts btw.
👇
Truly inspiring work.
99% of car interiors are terrible.
One can only hope this will set the bar for other manufacturers as the iPhone did for smartphones.
After many years of development, I’m excited to share the interior of the first electric Ferrari designed by LoveFrom. Tactile controls and digital interactions blend into one cohesive interface, shaped through deep collaboration across engineering, interaction, graphics, typography, sound, and industrial design. So incredibly proud of the thoughtfulness and care the team brought to every detail.
https://t.co/JZCleflfu7
A heartfelt appreciation and gratitude post to some incredible contributors and Rust developers that stepped up and brought Zed's git features to the next level.
The community stepped up beyond our imagination. You shipped great code, worked 1:1 with our team, and made this far more fun than us doing this alone. Cheers to shipping alongside your users. Especially if they're Zed users. ♥️🏆
See details in the blog post linked below. And big big thanks to the following contributors:
@0xbl4ze@amtoaer@amustaque97@BnJ25@cppcoffee@errmayank@heyxviraj@jafee201153@loricandre_@loricandre_@marcocondrache
“Alien mathematics” is my favorite thread this week: Consistent systems in “math space” that don’t map onto human intuition. Like wavelengths outside our visible spectrum, yet just as real. Wonder what kind of structures AIs “see” that we routinely miss…
🧵👇
Like @davidbessis and others, I think that Hinton is wrong. To explain why, let me tell you a brief story.
About a decade ago, in 2017, I developed an automated theorem-proving framework that was ultimately integrated into Mathematica (see: https://t.co/nGCIUk44TP) (1/15)
As I said: "if here's no catch": Details are super limited. No patents filed, it seems. Could be all just hype.
Lehtimäki in the video also appeared months ago in a vague "AGI" product presentation that doesn't look so good.🤷♂️
https://t.co/AtDb3PiYl9
If there's no catch, Finland's Donut Lab just made history.
A solid state battery that is cheaper, made with abundant materials, charges faster, with longer range, safer to use and available right now.
🤯
If there's no catch, Finland's Donut Lab just made history.
A solid state battery that is cheaper, made with abundant materials, charges faster, with longer range, safer to use and available right now.
🤯
Happy 2026 everyone!
I hope this is a year we choose to slow down in the ways that count and accelerate with wisdom. Listen more, assume less, and understand each other better. It takes time, there are no shortcuts and that's a feature, not a bug.
- Ready to use external context systems combining retrieval with temporal/entity graphs (Supermemory, Zep, etc.)
- Deeper architectural approaches like Google’s Titans + the MIRAS framework.
Is 2026 going to be the year of “AI memory”?
The more I use agentic AI, the more it seems that memory seems like the biggest issue. Makes sense, as agents work on complex multi-session tasks where context increases beyond what’s usually manageable.
Interesting to see various practical and theoretical approaches:
What we learned about memory in 2025
8 comprehensive resources:
▪️ Memory in the Age of AI Agents
▪️ When Will We Give AI True Memory? (interview with @EdoLiberty, founder & CEO @pinecone)
▪️ Why AI Intelligence is Nothing Without Visual Memory (interview with
@shawnshenjx, co-founder @memories_ai)
▪️ From Human Memory to AI Memory: A Survey on Memory Mechanisms in the Era of LLMs
▪️ Rethinking Memory in AI: Taxonomy, Operations, Topics, and Future Directions
▪️ Cognitive Memory in LLMs
▪️ MemOS: A Memory OS for AI System
▪️ MemEvolve: Meta-Evolution of Agent Memory Systems
Save the list and check this out for the links: https://t.co/sdlilMlYPo
@Simeon_Cps I mostly use Ghostty's quick terminal + OpenCode, when doing non-agentic development.
- Instant show/hide
- Notifies when done
- Covers ~50% of screen to view more output
- OpenCode + IDE shows changes in the repo
@karpathy I agree with the premise, but I wonder if agency is really something separate from intelligence or could be better seen as one of its (core) components.
BREAKING: Sugars essential for life have been found in pristine asteroid Bennu samples collected by NASA’s OSIRIS-REx spacecraft. Combined with previous detections of amino acids and nucleobases, we see that life’s ingredients were widespread throughout the solar system: https://t.co/Tb3HpwZG9J
More on the study led by Yoshihiro Furukawa of @TohokuUniPR⤵️
As a visual learner, I find these visualisations so much fun even if it's for a concept I learned a while ago.
Who knows, maybe in a few years #AI tools will generate them on demand as you learn.✨👇
I vibecoded this neural network visualization for my students and open sourced it.
It shows a simple MLP trained on MNIST handwritten digits at several training steps. The visualization is using @threejs and it comes with training code in @PyTorch .
Link + repo 👇