Remember that every news announcement from an AI lab is about $20,000 in product and marketing insights for free. When they provide a few bu…

Remember that every news announcement from an AI lab is about $20,000 in product and marketing insights for free. When they provide a few bullets of what a new model can do, it’s based off of the previous complaints and usage patterns of millions of customers. And we also need to remember that, in my opinion, Claude Code is in second place behind Codex right now. Read between the lines. This announcement tells me the following ⬇️ - Claude Code was annoying people by giving up early or jumping the gun on muddy assignments. I’ve seen Codex “give up” less, so that’s also competitive positioning. - Spreadsheets are a big enough use case that they get their own bullet! Excel and Google Sheets still likely hold the lion’s share…probably more Excel than GSheets if we look across all businesses in the US - 5.1 positioning (rather than .2 or .5) possibly to not oversell capabilities or spook the market - We’re still in a compute-constrained world (Fable 5.1 can be used for 50% of Max plan but need to see whether that’s on an hourly or weekly basis) - Verification is still a big ask from business users (I can confirm that this is one of the top 3 questions I get asked by business professionals)

Source: Allie K. Miller (X) | 2026-09-01

Related