Prediction / Featuring Eric Siu

Assessing the Prediction of AI Recursive Self-Improvement

Unresolved forecast: The speaker believes that AI systems will reach a level of recursive self-improvement within a year, surpassing current functionalities.

Editorial assessment
UNRESOLVED
Independent outcome evidence pending
Predicted: July 9, 2026

The Claim

From the recording

“I don't think we're there yet, where the the the system gets better on its own, right? We're not quite there yet. I I think we're going to get there within the next 12 months or so, but right now I think it's it's like, you know, a couple months ago, we're talking about Ralph loops, right? The Ralph Wiggum loops. Um, that was like a version to get to to slash goals. Now, we have slash goals. There's SL loop as well. And then event like it's going to get better and better over time. The models are getting better better over time and then eventually it's going to get into you're get get into recursive self-improvement.”

Eric Siu16:06

Unresolved forecast: The speaker believes that AI systems will reach a level of recursive self-improvement within a year, surpassing current functionalities.

What would settle this forecast?

The outcome is awaiting independent evidence. To assess it, identify the measure and time period in the original claim, then compare them with a dated observation covering that period.

Read the claim in context in How to Clone Your Best Employee Skills: Beat Prompts, Build a Skill Dojo, & Using Evals from Leveling Up with Eric Siu. The source records the argument as it was made.

Works Cited & Evidence

A source attribution identifies where a claim came from. It does not independently verify the claim. Inspect the original context and limitations. Editorial methodology · Corrections

1

How to Clone Your Best Employee Skills: Beat Prompts, Build a Skill Dojo, & Using Evals

primary source·Tier 3: Low-Authority Context·Leveling Up with Eric Siu·Jul 9, 2026

Primary source video

Disclosure: Social Signal Playbook prepared this archive entry with AI assistance.