Roon is a member of the technical staff at OpenAI, in a post from a couple of days ago. Plenty of insiders express similar views.
Two views on AI. On the left, an argument for superintelligence, soon, from “SITUATIONAL AWARENESS: The Decade Ahead” by former OpenAI employee Leopold Aschenbrenne. On the right, ChatGPT-4o stumbles on an easy problem, illustrated by prompt engineer Riley Goodside (who also write a lot about the upsides of AI as well)
I gave Claude 3.5 the paragraph above and asked for an illustration, here is what it came up with (I asked for a couple visual tweaks as well).
How to turn on artifacts (this should be free to try). You really should do this, then ask the AI to make an interactive explainer for a topic of your choice, to see what it can do.
You can play with the game here: https://claude.site/artifacts/945d2c95-e88c-4c55-b843-ccde906981c3
