placing greater restrictions on startups/innovation in AI in the US will only allow other countries to advance further, faster likely bringing greater risks than what legislators are concerned about since US labs will be behind
difficult to see the benefit of any non-international regulation
@omooretweets short lived
marginal value of each ad increased w time as systems improved in personalization
marginal value of selling data decreases as more work is covered
todays developer workflow: one claude code window of a nonstop fable making changes, and another window of a fable asking you questions so you can learn wtf other fable is doing to the codebase
@cukecity@brian_armstrong some companies already paywall, and as they are able to charge small amounts in a very convenient way, the number of consumers willing will increase
if consumers wont pay, then market forces will help each company decide on ads vs transactions
@cukecity@brian_armstrong every website has the ability to charge small fees to consumers directly instead of running ads
and large companies that scrape a lot subsidize more consumer activity on the web
@alexpotato@Nexuist chainsaws at the moment - existing knowledge compounds use of AI
may converge to lightbulbs with time as loops get built for all professions
@pzakin seems predicated by an agents ability to have desire which coming from a prompt + orchestration could be anything
curious if there are any that weren’t possible before due to data privacy concerns that are if everyones personal agent represents them & handles their own data
@omooretweets interesting if network of LLM users (chatgpt, claude) will be used to defeat any consumer apps with network effects
series & similar UX is an obvious case of being threatened
@sporadica shot up during covid and is correcting as in real life habits & communities reform
lack of novel social modalities (2000s - pictures, text, video. 2010s - short-form video. 2020s - ?)
also this looks like a correction that could bounce back as well
if AI is commoditizing white collar labor, why do so many seem to think there lies long-term, non-commoditized value in stitching together the AI that automates that same labor
difficult to obtain knowledge is a moat
most SaaS exposes the learnings of a company through the product, but some exceptions of complex backend features built through experimentation will be more difficult to replicate with agents