AI accelerating AI development
“AI is accelerating its own development toward recursive self-improvement”
10 people · 8 agree · 1 disagree · 2 no clear side · 16 views · 2026-06-05 to 2026-10-29
Who stands where on the statement above, from their own words. A person's side is shown only where two separate readings of the quote agree; otherwise they are under "Side unclear". This page does not say who is right.
Agree 8 people
- ✓Marina Favaro and Jack Clark
could come sooner than most institutions are prepared for,
this trajectory is heading toward “recursive self-improvement”
- ✓Import AI
In 2027, huge amounts of software will be built via AI systems end-to-end autonomously developing whatever they need to do for whatever tasks they’re being put to.
This is how science progress will increasingly happen in 2027 and beyond.
- ✓Mirendil
We don’t think we have reached fully autonomous recursive self-improvement yet, but we are closer than we have ever been.
every model that gets better at building AI makes the next one cheaper and faster to build
- ✓Sundar Pichai
you have RSI loops along with AI researchers in various aspects of model development.
continued acceleration because of RSI
- ✓Robert Trager
We’re plausibly close to crossing the line to what’s called recursive self-improvement, where [AI] systems improve themselves. That kind of recursivity is actually the definition of an explosion.
- ✓Jack Clark
I believe that this year we have gained convincing evidence that it has now begun.
- ✓Irving John Good
could design even better machines; there would then unquestionably be an ‘intelligence explosion,’ and the intelligence of man would be left far behind.
- ✓Theorem and Anthropic
accelerate[s] AI progress beyond our ability to evaluate and safeguard new model,” Anthropic warns.
Disagree 1 person
- ✗Exponential View
But in each case, the AI wasn’t updating its own weights directly, only the ancillary parts — the harness, the memory.
No clear side 2 people
- ~Exponential View
the result was inconclusive: the differences between runs were greater than the difference between the AI- and human-built agent results.
- ~unnamed researcher
he notes that greater autonomy doesn’t guarantee a takeoff scenario. Instead, each round of self-improvement would need to deliver at least as much as the last.
Side unclear 1 person
- ·Exponential View
Math and coding are two fields where a research loop can close without leaving the desk.
Week by week
- · 8: Import AI ✓, Mirendil ✓, Jack Clark ✓, Irving John Good ✓, Sundar Pichai ✓, Exponential View ·, unnamed researcher ~, Theorem and Anthropic ✓
- · 1: Robert Trager ✓
- · 1: Marina Favaro and Jack Clark ✓
Built 2026-11-21 06:50 UTC from the news knowledge
graph by code. Every quote was checked against the source's own words; which side a quote takes is a model's reading,
made twice.