soh3il

há 3 mesesdiscuss

Cool project! We've been building something in a similar space https://roundtable.now but took a different approach. Instead of polling models independently, ours runs sequential discussions where each model sees prior responses, then a moderator synthesizes everything into a single actionable output.

One thing we found is that the real value unlock is MCP integration. Instead of going to a separate UI to run debates, you can plug Roundtable directly into your coding agent, Claude Code, Cursor, VS Code Copilot, Gemini CLI, etc. and get multi-model council input without leaving your workflow.

soh3il

há 4 mesesdiscuss

Hey HN — I've been working on this for the past few months after reading the ICML 2024 debate paper and wanting to actually use it in my daily workflow.

The core insight: when Claude answers a question alone, it's confident. When Claude has to answer after reading Gemini's counterargument, it's more careful and often corrects itself. The adversarial pressure is the feature.

A few things that surprised me building this:

  1. Model diversity matters more than model quality. Three different models debating beats three instances of the best model.
  2. Sequential > parallel. Models that read prior arguments produce substantively different (better) responses than models answering blind.
  3. The moderator role is critical. Without synthesis, you just get three opinions. The moderator finds where they agree, where they disagree, and
   why.

 Would love feedback on the MCP integration experience — that's been the trickiest part to get right (resumable SSE streams, OAuth flow, structured tool outputs).

If Interested, hit me up and I share a 50% discount code with you.

soh3il

há 6 mesesdiscuss

Anyone else feels like they’re hitting a wall with pricing strategy lately?

I’ve been struggling with my 2026 roadmap for our SaaS product. We’re adding heavy AI automation, and it feels like charging "per seat" is literally punishing our users for being efficient. But every time I look at usage-based pricing, I get night terrors about churn and unpredictable MRR.

I couldn't get a straight answer from my mentors. half said "stick to seats for stability," the other half said "usage is the only way to survive AI."

Instead of spinning my wheels, I decided to run an experiment. I used a tool I’ve been working on called Roundtable.now where you can basically force different AI models (Claude, GPT, Gemini, Grok) to debate a specific problem.

I pitted them against each other. One model had to defend Seat-Based pricing, one had to argue for Pure Usage, and the third acted as a "Rational Skeptic" surfacing blind spots.

The Result: They actually found a "Third Way" I hadn't considered.

Instead of a binary choice between seats or usage, the models converged on a Hybrid "Platform + Capacity" Model. They essentially agreed that seat-based pricing is a "suicide mission" for AI-integrated SaaS because it creates a "tax on efficiency."

I’ve made the full debate log public here if you want to see the specific arguments and the "Hybrid Capacity" framework they built.

he most interesting part wasn't just the answer. it was the conflict. Watching the 'Analyst' argue for complexity while the 'Ideator' pushed for simplicity forced a synthesis that was much more robust than any single AI (or human) would have given me.

I’m curious, for those of you moving away from seats this year, what’s your "Value Metric"? Are you doing credits, tasks, or something else entirely?

Submissions

[untitled]

[untitled]

[untitled]

[untitled]

[untitled]

[untitled]

[untitled]

[untitled]

Can AIs Audit Each Other for Bias and Truth?

Is AI's Next Big Leap, Agentic AI, Set to Revolutionize or Ruin Tech Innovation?

comments