Ran 10,000+ queries through @Meta Muse via @NousResearch Hermes agents to see if AI recs = alpha for public equities
Core thesis: If 1B people use Muse, recurring patterns in product/service recs → identifiable edge in stocks
Setup:
- 11,567 Q&A records retained
- Mostly Muse Spark 1.3, some Wonder
- Dozens of sub-agents running concurrently
Key Findings:
1. Muse pushes frugality hard
Recommends free alternatives, fewer paid seats, repair over replace. Not just shilling premium products. Aligns with their cost-cutting ad narrative
2. But still recs premium when justified
Performance/reliability/convenience = willingness to pay more. It's situational, not dogmatic
3. AEO > SEO incoming
Companies will pay to get embedded in model parameters. Being the default rec = massive moat
4. Clear winners emerged:
- Fidelity: 58/70 brokerage scenarios
- Apple + Lenovo: 58/68 laptop recs
- Square: small biz payments default
- Best Buy: shopping destination
5. NOT auto-long those names yet
Popularity in model ≠ incremental earnings. Could just reflect current market share. Need to prove agents drive NEW profitable revenue beyond consensus estimates
Bigger picture:
AI agents could bifurcate markets — eliminate marginal spending while concentrating remainder into fewer winners. Default status in AI = new competitive advantage
Open sourcing full 11,567 record dataset. DM if you find investable angle
No clear trade yet but the framework is there
Core thesis: If 1B people use Muse, recurring patterns in product/service recs → identifiable edge in stocks
Setup:
- 11,567 Q&A records retained
- Mostly Muse Spark 1.3, some Wonder
- Dozens of sub-agents running concurrently
Key Findings:
1. Muse pushes frugality hard
Recommends free alternatives, fewer paid seats, repair over replace. Not just shilling premium products. Aligns with their cost-cutting ad narrative
2. But still recs premium when justified
Performance/reliability/convenience = willingness to pay more. It's situational, not dogmatic
3. AEO > SEO incoming
Companies will pay to get embedded in model parameters. Being the default rec = massive moat
4. Clear winners emerged:
- Fidelity: 58/70 brokerage scenarios
- Apple + Lenovo: 58/68 laptop recs
- Square: small biz payments default
- Best Buy: shopping destination
5. NOT auto-long those names yet
Popularity in model ≠ incremental earnings. Could just reflect current market share. Need to prove agents drive NEW profitable revenue beyond consensus estimates
Bigger picture:
AI agents could bifurcate markets — eliminate marginal spending while concentrating remainder into fewer winners. Default status in AI = new competitive advantage
Open sourcing full 11,567 record dataset. DM if you find investable angle
No clear trade yet but the framework is there