... plain text before you send it - is the worst of the four options tested. Two of the last three things I've written here were about reading the primary document instead of the coverage. This one is about a paper almost nobody covered, which is worse, because it changes a line of code a lot of us shipped this year without thinking about it. If your product has a "paste a URL" box, a link unfurler, an enrichment step ... ... competitor

... NBER paper tracks 500,000+ GitHub developers against their actual AI telemetry. The task-level gains are enormous and they mostly don't survive the trip to a shipped release. Across four app marketplaces: a sharp rise in new apps, no rise in total usage. Most of the AI-productivity argument you've read this year is two camps shouting single numbers at each other. There's a paper

... into three modes of inference: induction finds patterns in data; deduction derives conclusions from rules; and abduction proposes a new explanatory framework for a surprising result. Today's LLMs are strong at induction and rapidly improving at deduction. The paper argues that they still lack the abductive jump needed to originate new foundational hypotheses. Given Einstein's equivalence principle, a model might derive much of the later mathematics. The harder step is inventing that principle when observations are sparse ... ... world models that can act, observe consequences, and turn simulation into new formal hypotheses. LLMs may be excellent at climbing an existing hill. Whether they can jump to a hill that did not exist before remains an open question. Paper

... like infrastructure, not a nice-to-have. Following SafeLimit's progress! Comment from Jean-Olivier Brassard(@jean_olivier_brassard): @riya_pariyar  Why not replace granola with old school pen and paper ? Comment from Jean-Olivier Brassard(@jean_olivier_brassard): Claude, Claude design, Firefly, Apollo, Pen & Paper

... entire research workflow into one environment: searching millions of papers, building interactive knowledge graphs, analyzing research code, discovering connections across literature, identifying research gaps, and generating evidence-backed hypotheses. The biggest challenge I'm solving is making AI systems that can reason over evidence rather than just generate convincing answers. In scientific research, a claim is only valuable if it can be traced back to the underlying evidence: a paper

... actual insight, and it's bigger than it looks: making the notification say why it's waiting is what turns a ping into a triage. "Done" makes you tab back to find out what "done" means. "Needs approval vs finished vs

... writes the paper now. That part got easy. But then comes the actual submission part and suddenly it's 11pm on a Sunday and you're googling why your bibliography is broken again. Stuff researchers are still doing manually: fixing LaTeX that AI confidently got wrong fighting IEEE / Springer / Elsevier templates moving figures around for the 6th time converting a Word doc into LaTeX and regretting everything getting desk rejected because a margin was off AI made the writing faster ...

... started experimenting with is reading books, textbooks, or workbooks early in the morning, before touching my phone or any other electronics. So far, it seems to help improve my focus. How do you balance content consumption (especially digital vs analogue) to avoid shorter attention spans? And do you prefer e-books or physical books? Comment from Stan Kolotinskiy(@sk_uxpin): I'm reading almost exclusively e-books for several reasons: a) paper

... wanted Claude Code on my phone, so I built Clawd Phone, basically a mobile version of it. My phone has hundreds of PDFs and documents piled up: papers, books, manuals, screenshots, with no real way to search them. Now I just ask Claude things like find the paper