This part of the Karpathy/Dwarkesh interview was great. "Humans collapse during the course of their lives...we end up saying more and more of the same stuff, and the learning rates go down..." Many older people I admire have designed their lives to avoid this happening to them

what?

took me an hour to migrate openclaw over to minimax2.1 instead of claude 4.6 must be the user in this case. all it took was an aggressive prompt

decided to make an openclaw skill for job searching > uses theirstack api > api limited to 200 queries will probably go in and optimize the prompts for a cron job that runs every 12 hours one shotted the conversion from Frank's x-research repo github.com/zdeag/theirsta…

something obvious from the last couple of days > everyone is installing openclaw > no one knows what to automate integration is a completely different challenge if of it's self. the tooling is great but how does that actually impact numbers? i've talked to plenty of people in

was this deployed with Opus 4.6 or Sonnet 4.6?

memory lifecycles are such an interesting topic of discussion with ai agents. llm memory, retrieval-augmented generation, and context engineer often get confused with agent memory. while intrinsically connected, this very much differ in scope from: arxiv.org/pdf/2512.13564

decided that my claude code needed a new skill today > automatically dumps all session information: further questions, decisions, reasons into .md file > saves claude session id > dumps into obsidian vault

i can't be the only one thinking that aycd felt like pre-openclaw aycdjake

saw this gentle old man at chiptole sitting alone and second guessed sitting down with him. wonder what the conversation wouldve sparked.

have seen so many companies start hiring "ai user enablement & adoption specialists" lately something must be in the air

still think about this quote, have it written down on the whiteboard actually.

At some point you need to start understanding workflows and how to identify them. Applying agents to the problem is just half the problem. Most SMBs don't know where agents can be placed to see impactful improvements but fear the need for them. That's where the gap lies and will

great read newsletter.pragmaticengineer.com/p/how-claude-c…

as public investors, we've been kept out of the 'ai trade' for majority of the value > the beauty of the market lies in constant opportunity i've been struggling to figure out where to place my money after selling off most of my fintech exposure it's a good time to start

x.com/i/article/2025…