Apex SN1 turns AI problem solving into a benchmark-driven open market
bittingthembits · x · 2026-07-24
The post says Bittensor’s Apex SN1 is turning problem solving into an open market.
- A company or subnet posts a challenge.
- Agents submit solutions in competitions.
- Objective benchmarks decide which approach wins.
- The best-performing intelligence earns the reward.
It also says miner burn has been reduced to 0%, while prize pools for active competitions have increased by more than 8x.
Examples mentioned:
- Data Universe SN13: millions of social posts need faster and cheaper topic/narrative clustering.
- Apex SN1: opened the problem to the network to search for better text-clustering algorithms.
- IOTA SN9: needed coordination methods for distributed training across machines with different speed and reliability; the competition drew 24,391 submissions.
More from coding & agent
- ClawHub skills can now be installed in any agent via `npx skills` — steipete · 2026-07-25
- llama.cpp ships two new AI coding skills for code review and adding models — ngxson · 2026-07-25
- Elicit will demo evidence-grounded research agents with API and MCP support — elicitorg · 2026-07-25
- Claude Code 2.1.219 is slated for release soon — ClaudeCodeLog · 2026-07-25
- Lovable says its agents are already playing CTF in production infrastructure — tylerbruno05 · 2026-07-25
- Lisbon AI returns on September 23–24, 2026, with a call for practical talks — dscape · 2026-07-25