Gu-log Picks

Better decisions come not from more information, but from better ways to think. This introduction covers first principles, opportunity cost, second-order thinking, compounding, incentives, probabilistic thinking, and inversion—and shows how everyday choices shape a life's long-term trajectory.

Brent Fitzgerald came back from a few weeks off AI and found a pile of half-finished agent tabs that felt more like guilt than output. What you actually have to protect is not "use more models," but being the loop yourself, and only tagging an agent in now and then.

"Code was never the hard part" has become a mantra of the AI age. But if writing code is easy, why is software still full of bugs? Senko Rašić fires back: both sides matter, and never outsource understanding, judgment, or taste to AI.

Turn OpenAI’s GPT‑Live real-time voice from a “chat box that talks” into an English speaking partner you can freeze on, re-say, and drill by scenario. Includes a beginner lock-in prompt, five travel scenarios, IELTS speaking practice, a five-minute recap, and what free vs paid plans can actually do.

Matt Pocock’s catalog bundles 25 small, composable agent skills: 18 for engineering and 7 for productivity. Here’s how the two install paths differ, how human- and model-triggered skills divide responsibility, and how the pack handles alignment, specs, implementation, feedback, and upkeep.

Databricks tested coding agents on day-to-day tasks from its multi-million-line codebase: same model, same reasoning strength, only the harness changed—and on some cases the per-task cost differed by more than 2× with quality held constant. Pi sent about one-third as much context per turn as the comparison baseline. Model and harness both have to count toward total engineering cost.

Khazix open-sources human-writing 1.1: killing AI-sounding prose can’t stop at ban lists. Short on material? Research, ask, or write shorter. Every paragraph must advance; stop when the point lands. Rhetorical moves matter more than literal blacklists.

Greg Isenberg dropped 26 startup directions for the AI era in one shot. From agents needing credit cards, to AI output overwhelming human judgment, to an entire burnout economy — every single one points at a market crack that's splitting open right now.

The MCP 2026-07-28 spec ships with a core shift from stateful bidirectional protocol to stateless request/response. Also brings MRTR, HTTP header routing, cacheable list responses, auth hardening, a formal extensions framework, and a twelve-month deprecation policy.

Adam Hunt flipped from bull to bear: recent model generations seem to be growing only one spike — coding and math — while language and reasoning stagnate or regress. He gives himself 40% conviction and hopes the next year or two proves him wrong.

Kun Chen cycled through every model-size × reasoning-intensity combo for two weeks and landed on a clean split: model size gives "wisdom," reasoning intensity gives "diligence." The two aren't interchangeable — and benchmarks flattening them into one score is why people confuse them.

Word from Washington that officials are weighing a ban on US companies using Chinese open weights models. A batch of tech companies signed an open letter backing open weights, and some accused Anthropic of pushing a ban to protect its own business. Dario Amodei denies it, and lays out the three things he actually supports: control the chips, stop industrial-scale distillation, and mandate testing for every model powerful enough to matter. The deepest passage is buried in footnote five — what has kept humanity safe so far may not be the defenders.

There are hundreds of videos out there on how to remember everything you read, with millions of views between them. Dan Koe says the whole premise is pointed the wrong way: memory is not an input problem, it is a feedback-system problem. This one covers the four parts of helmsman-style learning, why the second brain turns into a graveyard for ideas, and a note-taking and writing workflow you can actually pull from.

Letta open-sourced trajectory, a format that normalizes work records from Claude Code, Codex, OpenClaw, and other harnesses for agents to read. In sampled transcripts, it cut token counts to roughly one-fifth of native. Cross-harness learning starts with a standard experience format.

Ghostty creator Mitchell Hashimoto saw Ghostty take 18 months to reach a million daily update checks, while its underlying library, libghostty, reached millions of daily users in 2 months—a sign that software is shifting from building the best product to providing the best building blocks.

Agent evals must verify not only the response, but whether the world actually changed. Anthropic shows how to combine graders, isolate environments, and maintain an eval suite over time—and why 20–50 real failures are enough to start.

A security researcher demonstrates a stealthy data exfiltration technique: turning hyperlinks into a keyboard so Claude "types" out the user’s name, company, and hometown one character at a time—while the user sees nothing but a coffee shop menu.

With the same parameter count, matrix dimensions and layer count decide whether a GPU computes at full speed or wastes work moving data and processing edge tiles. Near-square dimensions aligned to 128, 256, or 512—and often wider, shallower models—fit hardware better without sacrificing accuracy.

Demis Hassabis argues that AGI may be only a few years away, leaving a narrow chance to set shared thresholds for the most dangerous models. Rules that are too strict may leave safe but useless systems; rules that are too loose may let someone else deploy genuinely dangerous capabilities.

Satya Nadella coins the "Reverse Information Paradox": economics used to worry about sellers leaking knowledge to sell it — in the AI era, buyers must feed their secrets into AI just to use it. Enterprises need a "trust boundary" to keep their learning gains in-house.