@gabrielayuso.eth

This whole Clawdbot thing reminds me of the early consumer file and photo cloud storage days. There were countless startups and DIY solutions out there. Yet, at the end, the winners were the platform's built-in solutions (e.g. Google Drive, Google Photos, iCloud, Apple's Photos). Dropbox came the closest to be a breakout success in the consumer storage space yet they had to shift to Enterprise since the platform's built-in solutions became good enough. Like Steve Jobs said, they were a "feature, not a product". The AI space is playing out in a similar way. It's likely just a matter of time before the platforms themselves build these features in. It's not trivial for startups to do this because they don't have the models nor the user data, they're just acting like glue (hence the MCP trend that went nowhere). There is still room for Dropbox and Box-like startup successes but it's harder this time given the above. Anthropic and OpenAI would be the likely contenders IMO. There's also room for alternate clients like Superhuman but they're working with bigger constraints and are mainly a bet on being able to provider better UX than the platform they're building on top. I don't know much about Enterprise but there are likely strong opportunities there as well. Although players like Microsoft, Salesforce and GCP are likely making headway there.

2 replies

2 recasts

14 reactions