AI was supposed to make software cheaper. Incumbents used it to raise prices — then agentic coding evaporated the moat in 5 months
akumar · x · 2026-10-17
The author revisits his July 2025 thesis: SaaS incumbents (GSuite, Microsoft, Slack, Salesforce, LegalZoom) tacked on AI features to justify price hikes for 100% of users when only 1% would use them. He predicted disruption within 5 years — it took 5 months. The price hikes didn't stick and many firms now face existential risk from agentic coding. Software's 90% margins rested on a moat of 'nobody will rebuild a decade of features'; AI compressed that decade into months, so the moat evaporated rather than shrank. He cites BlackBerry's record revenue 3 years post-iPhone and Nokia's self-aware decline memos, and notes his own company Dragonfruit AI has cut prices yearly, concluding: companies charging everyone for AI features few use are living in BlackBerry's best year.
More from AGI Musings
- Neil Lawrence: AI creates two distinct cybersecurity problems, and leaders conflate them — lawrennd · 2026-10-18
- The 10,000-hour rule for applied AI: roughly 1 trillion tokens of practice — brandon_galang · 2026-10-18
- Palantir CEO: business now moves 5x faster; researcher warns of posthuman takeover — danfaggella · 2026-10-18
- 1966 news photo of teachers picketing calculator use resurfaces amid AI debates — MIT_CSAIL · 2026-10-18
- Murray Shanahan recommends SolidGoldMagikarp, a chronicle of the 2022-24 'LLM whisperer' era — mpshanahan · 2026-10-17
- FAIR researcher: cheap fast shallow thinking is replacing deep slow thinking — f_charton · 2026-10-17