Interconnects
robotic.skystack.xyz
Interconnects
@robotic.skystack.xyz
The cutting edge of AI, from inside the frontier AI labs, minus the hype. The border between high-level and technical thinking. Read by leadin...


This is an automated Substack Account of https://robotic.substack.com
Discover more/Create at: @skystack.xyz
Open models: Hot or Not with Nathan Lambert & Florian Brand • A recording from Nathan Lambert's live video
Open models: Hot or Not with Nathan Lambert & Florian Brand
A recording from Nathan Lambert's live video
www.interconnects.ai
December 18, 2025 at 8:35 PM
2025 Interconnects year in review • Three years writing every week about AI.
2025 Interconnects year in review
Three years writing every week about AI.
www.interconnects.ai
December 18, 2025 at 5:36 PM
2025 Open Models Year in Review • The first recap of a long year in the trenches of open models.
2025 Open Models Year in Review
The first recap of a long year in the trenches of open models.
www.interconnects.ai
December 14, 2025 at 8:35 PM
New Talk: Building Olmo 3 Think • Re-recording my NeurIPS talks in one mega-take.
New Talk: Building Olmo 3 Think
Re-recording my NeurIPS talks in one mega-take.
www.interconnects.ai
December 10, 2025 at 8:37 PM
Latest open artifacts (#16): Who's building models in the U.S., China's model release playbook, and a resurgence of truly open models • A month with SOTA releases with (truly) open model releases left and right.
Latest open artifacts (#16): Who's building models in the U.S., China's model release playbook, and a resurgence of truly open models
A month with SOTA releases with (truly) open model releases left and right.
www.interconnects.ai
November 24, 2025 at 5:40 PM
Olmo 3: America’s truly open reasoning models
Olmo 3: America’s truly open reasoning models
www.interconnects.ai
November 22, 2025 at 2:38 AM
Interview: Ant Group's open model ambitions • Launching a new open model series. Interconnects Interview #15.
Interview: Ant Group's open model ambitions
Launching a new open model series. Interconnects Interview #15.
www.interconnects.ai
November 18, 2025 at 10:40 AM
Opening the black box of character training • Some new research from me!
Opening the black box of character training
Some new research from me!
www.interconnects.ai
November 18, 2025 at 10:40 AM
5 Thoughts on Kimi K2 Thinking • Quick reactions to another fantastic open model from a rapidly rising Chinese lab.
5 Thoughts on Kimi K2 Thinking
Quick reactions to another fantastic open model from a rapidly rising Chinese lab.
www.interconnects.ai
November 18, 2025 at 10:40 AM
Burning out • An essay on overworking in AI.
Burning out
An essay on overworking in AI.
www.interconnects.ai
November 18, 2025 at 10:40 AM
How to scale RL • A bombshell paper from Meta and friends.
How to scale RL
A bombshell paper from Meta and friends.
www.interconnects.ai
November 18, 2025 at 10:40 AM
Latest open artifacts (#15): It’s Qwen's world and we get to live in it, on CAISI's report, & GPT-OSS update • After a quiet month, Qwen is back in full force.
Latest open artifacts (#15): It’s Qwen's world and we get to live in it, on CAISI's report, & GPT-OSS update
After a quiet month, Qwen is back in full force.
www.interconnects.ai
November 18, 2025 at 10:40 AM
The State of Open Models • Reflections on a crazy year in open models. (My talk at The Curve).
The State of Open Models
Reflections on a crazy year in open models. (My talk at The Curve).
www.interconnects.ai
November 18, 2025 at 10:40 AM
ChatGPT: The Agentic App • ChatGPT's long awaited move into user monetization and what it shows about the future of ChatGPT (and AI products writ large).
ChatGPT: The Agentic App
ChatGPT's long awaited move into user monetization and what it shows about the future of ChatGPT (and AI products writ large).
www.interconnects.ai
November 18, 2025 at 10:40 AM
Coding as the epicenter of AI progress and the path to general agents • GPT-5-Codex, adoption, denial, peak performance, and everyday gains.
Coding as the epicenter of AI progress and the path to general agents
GPT-5-Codex, adoption, denial, peak performance, and everyday gains.
www.interconnects.ai
November 18, 2025 at 10:40 AM
Latest open artifacts (#14): NVIDIA's rise, "Swiss & UAE DeepSeek," and a resurgence of open data • While Qwen takes some rest, others continue to fuel the open model space.
Latest open artifacts (#14): NVIDIA's rise, "Swiss & UAE DeepSeek," and a resurgence of open data
While Qwen takes some rest, others continue to fuel the open model space.
www.interconnects.ai
November 18, 2025 at 10:40 AM
On China's open source AI trajectory • Notes and predictions for the next phase of Chinese (open) AI models.
On China's open source AI trajectory
Notes and predictions for the next phase of Chinese (open) AI models.
www.interconnects.ai
November 18, 2025 at 10:40 AM
Ranking the Chinese Open Model Builders • From the obvious names to those to keep an eye on.
Ranking the Chinese Open Model Builders
From the obvious names to those to keep an eye on.
www.interconnects.ai
November 18, 2025 at 10:40 AM
Contra Dwarkesh on Continual Learning • Don't try to make your airplane too much like a bird.
Contra Dwarkesh on Continual Learning
Don't try to make your airplane too much like a bird.
www.interconnects.ai
November 18, 2025 at 10:40 AM
Latest open artifacts (#13): The abundance era of open models • Mostly thanks to Qwen, but now we're spoiled for choice and winds are shifting.
Latest open artifacts (#13): The abundance era of open models
Mostly thanks to Qwen, but now we're spoiled for choice and winds are shifting.
www.interconnects.ai
November 18, 2025 at 10:40 AM
What I've been reading (#2): More on Kimi K2, how to build a bad research center, Pretraining with RL, and sporks of AGI • A quiet summer is all you need.
What I've been reading (#2): More on Kimi K2, how to build a bad research center, Pretraining with RL, and sporks of AGI
A quiet summer is all you need.
www.interconnects.ai
November 18, 2025 at 10:40 AM
GPT-5 and the arc of progress • Overpromising will always lead to some sort of underdelivering, but what we're getting is still phenomenal.
GPT-5 and the arc of progress
Overpromising will always lead to some sort of underdelivering, but what we're getting is still phenomenal.
www.interconnects.ai
November 18, 2025 at 10:40 AM