Reid Hoffman's 2026 AI predictions
I really enjoyed this interview between Dan Shipper (of Every.to) and Reid Hoffman, co-founder of LinkedIn and general Silicon Valley genius. I think it's worth a look, even if you just read the Every write-up.
He makes 5 predictions that are worth your consideration, I think:
- Coding agents are the foundation for all AI tools
- AGI in 2025 will be one person with the capacity of a team
- Enterprises that sleep on AI will be left behind
- Discourse around AI might get uglier - even as the tools improve
- Biology will be the next 'language' AI learns
I think he's got a very good point with #1 as I'm increasingly seeing this in my own and colleague activities. For instance, I'm actually using Claude Code to get things done - as multiple assistants across multiple projects - not just for creating code. e.g. Earlier today, I asked it to find the details about an invoice I'm meant to be sending to a client. Claude Code pieced together about 25 different emails and made very, very good sense of them all. Total genius. It then gave me the template to send to the client – I wrote the email myself! So, for me, Claude Code is absolutely the foundation for everything else I'm using, already.
On point 2, I'm beginning to see that a lot with many of my consulting colleagues (across different industries) who are beginning to evidence the strength of many into what is often looking like super-human level activities and productivity!
I am particularly focused on item 3 on the list there, about enterprises asleep at the wheel. There is a LOT of this around. I am expecting to begin to start to see the difference between "AI get-its" and "AI-haven't-started" organisations.
There's a whole load of detail on Rhea's write-up of the podcast – and you can watch the whole thing right here:
