~7a woke naturally anyway
2:00m nuking still heats 16 oz tap-cold water past 140F; try 1:45
200mg caf while wrangling dogs, prepping cocoa et
clear and subfreezing cold - giving Columbia Trail trailhead scouting a miss.
29g protein bev
pipeline the new week's supplements and cocoa/soy.prot pods
The weekly vitamin/phytosterol pill minder calls out when you miss days - in this instance, Thursday and Saturday. The multivite and phytosterol protocols aren't all that sensitive to missed days; but creatine supplementation relies on a slow, consistent build-up process so a missed day is more disruptive.
* * *
Some months back I skimmed a few chapters of Forte's "Building a Second Brain". Core pitch bullet-points:
1. Information has become so prolific that human memory and attention are modern productivity limitations;
adopting a PIM (Personal Information Management system) helps offload retention and focus concerns.
2. PIMs bridge the gap between traditional note-taking, digestion and utilization of information
3. Concrete, practical workflow model to discipline and streamline the process of capture to action
| System | Primary Focus | Best For | Key Mechanism |
|---|---|---|---|
| GTD | Task & project execution | Getting things done, meeting deadlines | Next-action lists, weekly reviews |
| Zettelkasten | Knowledge creation & idea development | Research, writing, deep thinking | Atomic notes, bidirectional linking |
| BASB | Resource management for output | Creators, knowledge workers | PARA organization, progressive summarization |
DuckDuckGo launched "DuckAssist" LLM-backed (OpenAI/Anthropic) summaries to searches 2023-03-08.
Alphabet added "AI Overview" to general-public users' web searches 2024-05-14.
Ever since, LLM use-cases have captured executive attention and driven a tech pivot and that's been nothing short of transformative, with far-reaching and unpredictable global economic shockwaves from strategic mineral supply-chain to knowledge-worker labor pool. Sober analysis warns that the tech is overhyped and headed for a bubble-burst event - but like the proliferation of corporate penetration into the real-estate market (as owner / landlords, not just brokers and rehab "flipper" underwriters) in the wake of the 2008 subprime mortage securities meltdown ... what AI looks like and how it interacts with the general public post-crisis will probably (hopefully!) evolve past today's bland dark-pattern deflector chatbots and the colossally wasteful data-center boom to monetize billions of casual users asking the same 10,000 most-popular questions.
25 years ago, Spielberg completed Stanley Kubrick's last movie project, "A.I. Artificial Intelligence". Among the movies's multiple nonhuman characters, we have "Dr. Know" - an eerie premonition of what search engines have become with LLM summarization pushed to the forefront of user interface. I think this is far from the best and highest use of the tech, though - "Dr. Know", like DuckAssist or [Open]AI Overview, is just a layer of chatty presentation filter over data that might actually be better-presented elsewhere and delivered with much less logistical overhead and more pertinent/insightful context direct from Wikipedia, IMDB, Fandom, Goodreads etc.
Top Features to look for in a PIM include local-first storage for security and continuity assurance, interfaces for OCR and cloud storage (which in turn supports mobile access and multi-device sync). The fact that Alphabet's Gemini, Microsoft's Copilot, and other big players are using users' queries to iteratively refine their models, is a privacy concern, so we have 2 parallel directions emerging to keep private/proprietary information walled off: personal LLMs and privacy-forward services ala Swiss Proton's Lumo.
Naturally, with LLM service profit margin crashing to the freebie level, commercial PIM vendors are frantically bolting LLM service onto their existing wares. Notion's launch of "Notion AI", fitted to their established PIM product, is an iconic example.