In 2020, OpenAI showed that language models are few-shot learners. In 2022, ChatGPT was released. By 2025, Google DeepMind demonstrated that video models are zero-shot learners and reasoners. What’s next?
AGI
-
a16z Speedrun SR006: Last Day Applications Open for AI Startups
By
–
Last day to apply to a16z speedrun SR006. Applications close tonight (Sept 28 at 11:59PM PST).
— Bilawal Sidhu (@bilawalsidhu) 28 septembre 2025
I’m also writing scout checks – if you’re building something dope drop me a DM. Can’t reply to all but will review everything.
Prioritizing the areas below:pic.twitter.com/hwVX6joBQ3Last day to apply to a16z speedrun SR006. Applications close tonight (Sept 28 at 11:59PM PST). I’m also writing scout checks – if you’re building something dope drop me a DM. Can’t reply to all but will review everything. Prioritizing the areas below:
-
World Models: Essential Representations for AI Systems
By
–
world models are collections of representations; absolutely it matters what is represent but I think you are too dismissive of the need for world models. these are both important, overlapping questions.
-
Is Causality an Innate Concept in AI Systems?
By
–
i suspect the notion of causality is in some form innate
-
Brain Organization Precedes Experience in Learning Mechanisms
By
–
they don’t fully understand it, but a large fraction of brain organization precedes prior to or independently of experience, and underwrite the various kinds of learning mechanisms that different creatures have.
-
Clarification on Starship and AGI Feasibility Claims
By
–
show me where i said either that starship or AGi was impossible. oh wait i never said either. you just made both up
-
Chain of Thought vs True Reasoning in LLM Next Token Prediction
By
–
Feels about equivalent to chain of thought in next token guessing LLMs being used for "reasoning" to me
-
AGI Jobs Edge-of-Knowledge Sprint Implementation Pipeline
By
–
[ AGI Jobs ] Edge-of-Knowledge Sprint Plan: "Implementing the Planner → Simulator → Runner Pipeline" https://
docs.google.com/document/d/1PR
YyWxL-JNNpuYTT4gTXmMpKcPyW6Xzz/edit?usp=sharing&ouid=115006596614780570930&rtpof=true&sd=true
… GitHub: https://
github.com/MontrealAI/AGI
Jobsv0
…) $AGIALPHA #AGIALPHA #AGIJobs -
The Bitter Lesson’s Bitter Lesson: AI Learning from Human Knowledge
By
–
.
@iamtrask has a terrific new blog post, the Bitter Lesson's Bitter Lesson, where he concludes: "Nature spent billions of years learning not to start over—and the future belongs to AI systems that can inherit vast human knowledge while maintaining robust capabilities for novel