StarZero All shows Try the podcast clipper →

Jun 9, 2026 · 22m · startup-ideas

WTF Is an "AI Agent Loop"? Genius or Hype?

Ross Mike · 16m spoken Greg Isenberg · 4m spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

Host Greg Isenberg and guest Ross Mike critically evaluate the hype surrounding AI agentic loops, contrasting the costly pitfalls of unconstrained autonomous loops with practical, deterministic use cases like automated code review.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Greg holds 22.3% of the talking time here. How this is scored →

Greg as informed peer 3.4 Guest teaching 5.4 Guest disagreement 3.8 Greg pushing back 0.8
05100:0010:0020:000:35–4:36 · Greg as informed peer 3/10 Setting Expectations: Debunking Hype vs Practical Reality Ross sets an aggressive anti-hype thesis against agentic loops while introducing the mechanics of human-in-the-loop workflows. Greg acts primarily as a supportive moderator, summarizing Ross's opening thesis cleanly.4:38–7:59 · Greg as informed peer 2/10 Sponsor Spotlight: Automated Code Reviews with CodeRabbit Following an ad read, Ross uses an analogy of a rogue developer to explain why unconstrained agent loops fail and waste immense token budgets. Greg is largely absent from the core discussion during this breakdown.8:00–12:58 · Greg as informed peer 4/10 Slash Goal, Token Consumption, and the 'Slop Machine' Ross details why trending tools like Slash Goal fail on nuanced apps and burn capital, citing benchmark tests. Greg contributes by crystallizing the issue into the term 'slop machine.'12:58–18:18 · Greg as informed peer 1/10 Practical Application: Building a Code Review Loop with Greptile and Cursor Ross delivers an extended technical walkthrough of a functional code review loop using Cursor and Greptile, showing its strict constraints. Greg remains silent while Ross provides an in-depth tutorial.18:20–21:51 · Greg as informed peer 7/10 Binary Systems vs Creative Startup Execution Greg demonstrates strong domain insight by analyzing why startup building requires iterative human pivots unlike binary deterministic tasks. Ross actively agrees and validates Greg's framing.0:35–4:36 · Guest teaching 5/10 Setting Expectations: Debunking Hype vs Practical Reality Ross sets an aggressive anti-hype thesis against agentic loops while introducing the mechanics of human-in-the-loop workflows. Greg acts primarily as a supportive moderator, summarizing Ross's opening thesis cleanly.4:38–7:59 · Guest teaching 6/10 Sponsor Spotlight: Automated Code Reviews with CodeRabbit Following an ad read, Ross uses an analogy of a rogue developer to explain why unconstrained agent loops fail and waste immense token budgets. Greg is largely absent from the core discussion during this breakdown.8:00–12:58 · Guest teaching 6/10 Slash Goal, Token Consumption, and the 'Slop Machine' Ross details why trending tools like Slash Goal fail on nuanced apps and burn capital, citing benchmark tests. Greg contributes by crystallizing the issue into the term 'slop machine.'12:58–18:18 · Guest teaching 8/10 Practical Application: Building a Code Review Loop with Greptile and Cursor Ross delivers an extended technical walkthrough of a functional code review loop using Cursor and Greptile, showing its strict constraints. Greg remains silent while Ross provides an in-depth tutorial.18:20–21:51 · Guest teaching 2/10 Binary Systems vs Creative Startup Execution Greg demonstrates strong domain insight by analyzing why startup building requires iterative human pivots unlike binary deterministic tasks. Ross actively agrees and validates Greg's framing.0:35–4:36 · Guest disagreement 4/10 Setting Expectations: Debunking Hype vs Practical Reality Ross sets an aggressive anti-hype thesis against agentic loops while introducing the mechanics of human-in-the-loop workflows. Greg acts primarily as a supportive moderator, summarizing Ross's opening thesis cleanly.4:38–7:59 · Guest disagreement 5/10 Sponsor Spotlight: Automated Code Reviews with CodeRabbit Following an ad read, Ross uses an analogy of a rogue developer to explain why unconstrained agent loops fail and waste immense token budgets. Greg is largely absent from the core discussion during this breakdown.8:00–12:58 · Guest disagreement 5/10 Slash Goal, Token Consumption, and the 'Slop Machine' Ross details why trending tools like Slash Goal fail on nuanced apps and burn capital, citing benchmark tests. Greg contributes by crystallizing the issue into the term 'slop machine.'12:58–18:18 · Guest disagreement 3/10 Practical Application: Building a Code Review Loop with Greptile and Cursor Ross delivers an extended technical walkthrough of a functional code review loop using Cursor and Greptile, showing its strict constraints. Greg remains silent while Ross provides an in-depth tutorial.18:20–21:51 · Guest disagreement 2/10 Binary Systems vs Creative Startup Execution Greg demonstrates strong domain insight by analyzing why startup building requires iterative human pivots unlike binary deterministic tasks. Ross actively agrees and validates Greg's framing.0:35–4:36 · Greg pushing back 1/10 Setting Expectations: Debunking Hype vs Practical Reality Ross sets an aggressive anti-hype thesis against agentic loops while introducing the mechanics of human-in-the-loop workflows. Greg acts primarily as a supportive moderator, summarizing Ross's opening thesis cleanly.4:38–7:59 · Greg pushing back 0/10 Sponsor Spotlight: Automated Code Reviews with CodeRabbit Following an ad read, Ross uses an analogy of a rogue developer to explain why unconstrained agent loops fail and waste immense token budgets. Greg is largely absent from the core discussion during this breakdown.8:00–12:58 · Greg pushing back 1/10 Slash Goal, Token Consumption, and the 'Slop Machine' Ross details why trending tools like Slash Goal fail on nuanced apps and burn capital, citing benchmark tests. Greg contributes by crystallizing the issue into the term 'slop machine.'12:58–18:18 · Greg pushing back 0/10 Practical Application: Building a Code Review Loop with Greptile and Cursor Ross delivers an extended technical walkthrough of a functional code review loop using Cursor and Greptile, showing its strict constraints. Greg remains silent while Ross provides an in-depth tutorial.18:20–21:51 · Greg pushing back 2/10 Binary Systems vs Creative Startup Execution Greg demonstrates strong domain insight by analyzing why startup building requires iterative human pivots unlike binary deterministic tasks. Ross actively agrees and validates Greg's framing.

speaking balance: gold is Greg, purple is the guest (3 minute bins)

0:00 · Greg 34.2% · guest 65.8%0:00 · Greg 34.2% · guest 65.8%3:00 · Greg 28.3% · guest 71.7%3:00 · Greg 28.3% · guest 71.7%6:00 · Greg 4.6% · guest 95.4%6:00 · Greg 4.6% · guest 95.4%9:00 · Greg 4.9% · guest 95.1%9:00 · Greg 4.9% · guest 95.1%12:00 · Greg 0% · guest 100%12:00 · Greg 0% · guest 100%15:00 · Greg 0% · guest 100%15:00 · Greg 0% · guest 100%18:00 · Greg 75.3% · guest 24.7%18:00 · Greg 75.3% · guest 24.7%21:00 · Greg 42.4% · guest 57.6%21:00 · Greg 42.4% · guest 57.6%
Sharpest disagreement ▶ 7:15 Calling trending loops a catastrophe

Ross forcefully dismisses the trending narrative around agent loops, arguing that without infinite token budgets, autonomous loops are completely counterproductive.

 Hardest push from Greg ▶ 19:55 Defending Boris and Peter's forward-looking research

Greg gently pushes back against completely writing off agent loop creators, noting that while current startup builders shouldn't use them, frontier researchers are paving the future.

 Biggest teaching moment ▶ 5:30 The autonomous developer assumption flaw

Ross educates on how autonomous agents inevitably hallucinate or misalign when building full software stacks from static PRD files.

 Greg holds their own ▶ 18:20 Differentiating binary tasks from creative startup iteration

Greg demonstrates deep product expertise by framing startup creation as non-binary road trips requiring intermediate user feedback rather than locked-in automated loops.
the scores for every segment, with the reasoning behind each
ChapterTopicGreg as informed peerGuest teachingGuest disagreementGreg pushing backWhy
Setting Expectations: Debunking Hype vs Practical Reality 3541 Ross sets an aggressive anti-hype thesis against agentic loops while introducing the mechanics of human-in-the-loop workflows. Greg acts primarily as a supportive moderator, summarizing Ross's opening thesis cleanly.
Sponsor Spotlight: Automated Code Reviews with CodeRabbit 2650 Following an ad read, Ross uses an analogy of a rogue developer to explain why unconstrained agent loops fail and waste immense token budgets. Greg is largely absent from the core discussion during this breakdown.
Slash Goal, Token Consumption, and the 'Slop Machine' 4651 Ross details why trending tools like Slash Goal fail on nuanced apps and burn capital, citing benchmark tests. Greg contributes by crystallizing the issue into the term 'slop machine.'
Practical Application: Building a Code Review Loop with Greptile and Cursor 1830 Ross delivers an extended technical walkthrough of a functional code review loop using Cursor and Greptile, showing its strict constraints. Greg remains silent while Ross provides an in-depth tutorial.
Binary Systems vs Creative Startup Execution 7222 Greg demonstrates strong domain insight by analyzing why startup building requires iterative human pivots unlike binary deterministic tasks. Ross actively agrees and validates Greg's framing.

Statements from this episode (8)

Opinion
Mike: AI agent loops are a mistake without money to burn
“You're gonna understand why people are fanning out about it, and you're gonna understand why it is a terrible mistake, and unless you have money to burn, that you are not to do it.”
Ross Mike Jun 9, 2026 ▶ 0:44
Insight
Mike: Autonomous AI agents usually guess wrong and burn money
“And believe me, when you give the agent the floor to give assumptions, most of the time it's going to get it wrong, but not only is it going to get it wrong, it's going to burn a lot of money.”
Ross Mike Jun 9, 2026 ▶ 6:38
Insight
Mike: AI can replicate sauce but cannot create sauce
“When you and I are trying to use AI to build something meaningful, I hundred percent stand in the fact that the human still needs to be in the loop. AI can replicate sauce. It can't create sauce.”
Ross Mike Jun 9, 2026 ▶ 10:59
Assertion Supported
Mike: Peter burned $1.3M in AI tokens in one month on loops
“The one argument I'll fight back with is this is going to burn a lot of tokens, and if you don't believe me, all you have to do is look at Peter's tweet, where in one month he burnt 1.3 million dollars worth of tokens.”
Ross Mike Jun 9, 2026 ▶ 12:29
Disclosure
Mike refuses to ship code scoring under 4/5 on AI review
“The mental model I now have is I will not push anything to production, meaning I will not allow code to go live unless the score is greater than four out of five, right?”
Ross Mike Jun 9, 2026 ▶ 14:18
Insight
Mike: AI agent loops only work in constrained feedback environments
“The only place a loop makes sense is in a very confined, constrained process with a very fixed feedback loop, a very defined feedback loop, and that's in code review.”
Ross Mike Jun 9, 2026 ▶ 16:43
Insight
Mike: AI review loops break down on PRs exceeding 1,000 lines
“Anytime I push over 1000 lines of code, One K lines of code. Like if the code that it has to review is more than 1000 lines, I can almost never get a five out of five because it's too much code for the agent to fully review and contextualize and understand.”
Ross Mike Jun 9, 2026 ▶ 17:03
Insight
Isenberg: AI agent loops only work for binary, non-creative tasks
“I think where, where the output is binary, meaning black or white with no creativity, there is a room for loops.”
Greg Isenberg Jun 9, 2026 ▶ 19:55
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 300 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.

Try the podcast clipper Learn more