I recently went to a playground with my kid, there were a few other fathers there. We exchanged understanding looks and then I realised I have absolutely nothing, zero, nada, in my head to tell them outside of work stuff. I panicked, so I asked if they are using agentic coding, all of them were, we talked about work stuff. There's no escape :)

Got terminals to survive across Kolu restarts & deploys. So ... persistent terminals, like tmux. Powered by Kaval: kolu.dev/kaval/ If you want to test: github.com/juspay/kolu/pull/…

Welcome to Kolu Atlas. Entirely LLM managed. Superseded both "HTML plans" and GitHub issues for me. Tracked permanently in Git, and uses Astro. You can view it here: htmlpreview.github.io/?raw.g…

let your agents iterate on the actual app in the browser, and post 'screenshot evidence' in your PRs kolu.dev/blog/nix-chrome-dev…

Appreciation is also anti-"timeline" It can only be of what is happening now. As what is happening right now / here is all that was / is / will be happening, there is nothing else (x.com/sridca/status/20445998…) to do than enjoy & appreciate it. If you want something, do it now.

And today has been even better than that. Because, my experience has been the closest it has ever been to apperception — "surging perpetually on the crest of the ongoing wave of this moment in eternal time" (even instinctual passions, while still-born, segued into sensuosity).

Actualism For Dummies (created by Fable ultracode) actualismfordummies.pages.de…

NEW: malware developers added nuclear & biological weapons text to to their spyware. Goal? To trigger LLM safety refusals... so that their spyware wouldn't be analyzed by an AI security scanner. Cleanest practical example I can think of for why over-indexing on first order safety alignment is risky. When closed (and open) models ship with aggressive refusals, they will be sprinkled with second-order blindspots that attackers will discover...and exploit. We are only in the earliest days of attackers leveraging these features, and it wouldn't surprise me if users systems that need to handle complex cybersecurity issues demand that models be less safety-blunted. In the weeds: @SocketSecurity's post also shows why intention matters in how you design a malware analysis pipeline to avoid prompt manipulation. H/T to colleagues that shared this with me socket.dev/blog/mini-shai-hu…