AI Leader, Researcher, and Builder
Curating AI content every week, so you don’t have to.
Episodes worth the commute, with the reason I picked each one
The ones I keep going back to, not the ones everyone is posting about
Papers worth the hour it takes to read them properly
The week in AI video, minus the thumbnails shouting at you
What to put on for architecture, flow, a sprint, a bad bug, or 2am
What practitioners are actually arguing about this week, and who is making sense
What I think the week added up to
A new piece every day, written from what the field published that morning
Where separate communities landed on the same thing without talking to each other
The releases people cannot stop bringing up
Follow one subject through a paper, an episode, a book and the analysis
Keeping up was costing me about twenty hours a week, and most of those hours were wasted on things that turned out not to matter. So I built something to do the hunting.
It works the way you pick a restaurant. One friend recommending a place is a suggestion. Four friends who have never met, all mentioning it in the same week, is a signal — and the fact that they never spoke to each other is precisely what makes it one. This site watches communities that have no connection to one another and keeps only what several of them arrive at separately. Whatever survives is scored, and every threshold that score depends on is printed further down this page, with the arithmetic shown on the card so you can check it yourself.
A typical run through the video candidates weighs around two hundred of them and publishes five. Some runs add nothing at all — the shelf goes short, or keeps what it had, rather than padding with something that did not earn a slot. Below is exactly how that decision gets made.
Read from this site’s own API while the page loaded, not typed into the copy. If they are bad today, they are bad on the page.
Podcasts, papers, books, video, music and short written insights. Automated picks pass selection checks; Brendon’s pinned books are identified on their cards. Reader suggestions can enter discovery and face the same checks as other candidates.
Every whitepaper opens with a short spoken lead-in: what this paper claims, and whether it deserves your next forty minutes. Then the available paper text, read aloud. Use the lead-in to decide whether to continue, or open the source to inspect the original figures and claims.
It is a synthetic voice. A good one. You will still know.
The chat carries my interests and how I read this industry, so you can argue with the reasoning instead of reading a verdict. Two per visitor per day, because each one calls a real model and those cost money.
It speaks for me. It is not me, and it will be confidently wrong sometimes. So am I.
And nobody approves the front page. Not me either. Most mornings I find out what is on it the same way you do. So the next fair question is: what exactly is deciding, and why should you trust it? That is the rest of this page.
Nothing gets in because one person liked it. Separate crowds have to arrive at it independently — and the site has to be able to show that they did.
In words: every mention counts for something, a trusted source counts for more, and a mention loses value the older it gets. Add up what is left. That is the whole idea — the line below is just that sentence written precisely enough to run.
value(s) = authority × e−ageDays / 7 × (0.5 + 0.5 · strength)score = min(10, Σ value(s) / 2.5) over distinct groups
Calibrated so that one fresh, top-authority sighting reaches about 3.6 and three of them reach 10. A single source cannot climb into corroborated territory alone, however loudly it shouts.
Exact fingerprints require identical normalised text, so a formal paper title and a community’s paraphrase never matched — 13 of 1,080 signals converged. Fuzzy title matching closed that gap deterministically, at no API cost. The threshold was measured, then tightened when a test found it too loose.
Jaccard overlap of content tokens. Genuine and false pairs separated cleanly with nothing in between; 0.5 sat inside the gap but hugged the wrong edge. Hover any point for the pair it came from.
A fuzzy match contributes no keys, so it can never seed another round of matching. Chained approximate matches drift, and a false merge here manufactures convergence — which is the failure this whole page exists to avoid. A containment rule was tried and rejected: it merged a model announcement with a benchmark analysis of the same release.
Eight steps take something from “somebody mentioned it” to “it is on the shelf” — and the last step changes the first. Automated picks follow this pipeline. Pinned books and original writing have their own selection and review paths.
Seven feeds are polled independently: research paper listings, a link-voting board, subreddits, named editors writing newsletters, developer platforms, and two social timelines. Each sighting is recorded with who saw it, how strongly, and when — nothing is judged yet.
The same paper might arrive as an arXiv link, a video about it, and a code repository. Each mention is reduced to a handful of identifiers — arxiv:2601.01234, yt:…, gh:owner/repo, a tidied-up URL, a tidied-up title — so the three can be recognised as one. A strong identifier outranks a bare URL, and addresses like youtube.com/watch are refused outright: on their own they are the same string for every video on earth.
Anything sharing an identifier is merged. Titles that are obviously the same thing worded differently get exactly one chance to join — and a match made that way contributes no identifiers of its own, so it can never set off a further merge. Chains of approximate matches drift, and a wrong merge here would manufacture the very agreement this site exists to detect.
One community arriving over three transports collapses to one vote. Sources declared non-independent collapse together. Publication venues and first-party announcements are shown and never counted. An undeclared source contributes zero — the registry fails closed.
Course, listicle and clickbait title shapes are blocked before anything is scored. An unregistered source must show exceptional organic traction instead. Commentary that speculates about content it never saw is blocked after generation. All deterministic, all in code.
The model grades only the dimensions the prompt actually defines. Its quality grade becomes the starting point; corroboration, standing, freshness and measured traction are bonuses that can only push a score up, never drag one down. Nothing below 8 out of 10 is published as though it cleared the bar. In a thin week the best of what is left can appear rather than leaving a shelf broken, marked 'best this week' and showing its real score — the number on the card is never adjusted to fit.
A single module copies a named list of files from the private side to the public snapshot, stripping anything personal on the way. It is the only thing that writes published content, and a test fails the build if another module ever starts. Every public route — and the chat — reads only the snapshot, never the private side.
Every published shelf is graded and the grade is published unedited, good or bad. Separately, each source is re-measured on how often its picks were independently confirmed. A strong record can raise a source’s standing, one step per review. A weak one cannot lower it — “nobody else mentioned its picks” is a fact about everybody else, not about the source. Only going quiet for weeks lowers a tier.
The same journey as numbers, from the most recent video run — read live from /api/content/youtube while this page loaded. Five points where the count actually drops; each bar is what survived the one above it.
08 → 01. What the field independently confirmed feeds back into how much each source is worth on the next sweep. Nobody edits that number by hand — and it is deliberately a ratchet upward, because a source going unmentioned says more about the room than about the source.
Current automated picks use a quality anchor with measured bonuses. Some books, music and older records retain earlier rubrics; each card uses its stored calculation. A thin shelf may include a labeled below-band pick at its actual score.
A weighted average diluted strong items: a genuinely good video whose freshness and engagement were merely average got dragged down to about 5, while shelves whose displayed score is essentially the quality grade sat at 8–9. Same axis, incomparable distributions.
The whole bonus budget is the top +2.00 of the scale. Everything under it is the grade. That is why a well-corroborated mediocre item still does not ship: from a grade of 6 the bonuses top out at 8.0 exactly, and they are almost never all maxed at once.
Every component is a bonus, so nothing gets gutted for lacking views it has not had time to bank — and measured corroboration earns a positive lift until the score reaches its ceiling. Each bonus names where it comes from: corroboration from independent communities (decayed), authority from the source’s earned tier, engagement from real measured traction only, freshness computed from the publication date and never asked of the model.
The model is asked one thing: how good is this, against a rubric it is shown. It is not asked how fresh something is — a publication date is arithmetic, and a model guessing at it from a title is a made-up number in a slot that looks measured. Freshness, corroboration and engagement are all computed in code from things that were actually counted.
The division is the point. A model is good at judgement and bad at arithmetic, so it is given the judgement and none of the arithmetic. The rubric and the JSON it is asked to return share one definition, so the question and the answer format cannot drift apart — they did once, and every sub-score on three video cards came back as the same number wearing four labels.
A model can be talked into anything by a well-produced title. Some judgements are therefore not the model’s to make: they are regular expressions and integer comparisons that run before scoring, and again after generation.
Course, listicle and clickbait shapes are a category error, not a quality judgement. A well-produced “Zero to Hero Full Course” can legitimately score 7.9 on depth and still be study material rather than curation. Two classes: most block unconditionally, and the ambiguous ones (marked) block only for sources without earned authority — a Dwarkesh interview titled “Part 2” is exactly what this engine is for.
Six of nineteen shown. the marked one blocks only for sources without earned standing. Every block is logged with the pattern that caused it, so the record explains why a title died rather than just that it did.
Unregistered video sources face authority-or-traction checks. An authoritative guest can qualify an interview. A missing view count is never treated as a measured zero.
“Agentic AI Full Course with LangGraph … (Zero to Hero) — Part 1”, unknown channel, 15,000 views, published to the video shelf at 7.9. The rubric was not wrong about its production values. The rubric was the wrong instrument.
Commentary is published in Brendon’s voice, as his read of an item. A podcast curator works from publisher descriptions and available transcript excerpts. When the model wants to characterise the episode anyway, it hedges and keeps going. Two published lines, live on 2026-07-30:
“This episode likely cuts through the hype better than most…”
“The conversation with Sarah Guo likely cuts past the usual AV hype…”
A reader cannot tell that apart from a genuine assessment — it is a confident claim the evidence does not support, carrying a person’s name. Nine patterns now block that shape after generation, alongside three other families that catch invented biography, narration of the rubric, and source name-dropping — all applied to the published prose after generation. Having an opinion is the job; pretending to have watched something is not.
A judge that had no hand in choosing the content reads every shelf and publishes the mark, good or bad.
Three separate model passes — lenses — read each finished shelf and score it 1–10. The questions are not generic, because a shelf can only be judged against what it is for. Focus music graded on novelty would fail for being repetitive, which is precisely what focus music should be.
Yes — a model is grading work a model helped produce. Two things stop that being a rubber stamp.
The judge is not the author. It runs on a model from a different lab than the one that scored the items or wrote the commentary, and the model that actually answered is recorded on every grade. When the two collide anyway, the grade is withheld rather than published — a missing mark is a gap someone can chase, a flattering one is a lie with a footnote. The day this separation was introduced, insights fell from 9 to 3.3 and discourse from 9 to 3.7 on unchanged content: the self-graded numbers had been running about five points generous.
A judge cannot grade what it was not shown. Asked to score something it cannot see, a model invents a number rather than admitting it. So every lens declares the evidence it needs — the summary, the date, the source, the duration — and a test renders each shelf’s real judge input and fails the build if a lens was not shown what it requires. A freshness lens once scored four shelves it had never been given a date for. It cannot happen silently again.
Three loops revise the machine's own inputs from its own record. The counts below are read live from /api/content/self-improvement — a loop that stops running shows a frozen number here rather than prose that keeps claiming it works.
The site publishes an agent card at /.well-known/agent-card.json and answers JSON-RPC at /a2a. Every skill it advertises is derived from the same table that serves it, so it cannot advertise something it does not do.
reading /api/content/audit…
The weekly grade is the slow check. The fast one runs on every code change: a set of labelled examples — including the real failures that caused this engine to be built — is pushed through the actual gates and the actual scoring, using recorded model answers so the result is identical every time and costs nothing to run. Three things must hold, or the change does not ship.
These are the raw feeds the page reads from. Open any of them.
/api/content/pulse
what is being tracked right now, which sources count as independent of each other, and why — written by the same code that does the counting, so it cannot describe a rule it is not running
/api/content/audit
this week’s grades, and the criticism behind each one, in the grader’s own wording
/api/content/ledger
the track record of what was published and what happened next
/api/content/topics
a deliberately weaker claim: what is being discussed, not cross-verified
That was the whole trade. The twenty hours a week went into feeds, tabs and half-finished episodes, and almost all of it was spent finding out what not to read. That part now happens overnight, in code, against thresholds printed on this page. What is left on the shelves is short on purpose — some days a shelf adds nothing, and would rather stay short than pad itself.
The simplest way to use it is to let it come to you. Every Monday at 7am ET: what cleared the bar that week, each item linked to its own page here with the score breakdown attached. No digest of a digest, and one click to leave. If you would rather argue with it, the chat takes two questions a day and I read what comes through.
