ChatGPT hit 30% adoption in roughly 2 years, faster than PCs, the web, or smartphones. But Benedict Evans argues that comparison is rigged: ChatGPT is a free website with wall-to-wall media coverage, no hardware purchase required, no broadband rollout to wait for. The speed is real. So is the asterisk.
The real number to watch is DAU, and Evans points out that Sam Altman never gives it. Altman's headline figure is 1 billion weekly active users globally. Evans reads that omission as a signal: somewhere between 5% and 15% of people find a use for this daily, while at least twice as many know how it works and still only reach for it once a week. Social media taught us that weekly and monthly active figures are vanity metrics. If ChatGPT were truly working for most users, Altman would lead with DAU.
Evans lays out 3 possible explanations: a time problem, where habits haven't formed yet and the S-curve is still climbing; a product problem, where the technology needs its iPhone moment to crystallize, with Evans literally paging Jony Ive; or a UX problem, where the chatbot interface is only right for a subset of users and most people will encounter this technology as embedded features inside other products. He doesn't pick a winner. Read the original for why he thinks the 3G 'killer app' question was the wrong frame then, and might be again now.
[READ ORIGINAL →]