The Short Version turns podcasts into what was actually said. Every
weekday we take the shows worth knowing about — macro, markets, tech,
power — and boil each episode down to the argument, the takeaways, and
the calls, if anyone was brave enough to make them. An hour of someone
talking becomes a few minutes of you reading. You listen to the episodes
that earn it and skip the rest with a clear conscience.
How it works: subscribers get one email every weekday around 6:30am ET —
the morning's index plus the takeaways from each episode. The full reads
live here on the site: the argument, the reasoning, what has to be true
for the person to be right. For now, we do things in categories. You
pick which ones land in your morning email — the links at the bottom of
every digest handle that, plus unsubscribe if it comes to that. No
account settings to dig through, and no passwords anywhere: the email
itself logs you in. Your inbox is your identity around here.
Want a podcast covered? Reply to any digest or email
requests@givemetheshortversion.com.
We'll decide if it makes the cut. "Hey, why do you get to
decide?" Because this thing is free for now. Send money and your opinion
will start carrying real weight around here.
Fair warning: this thing is young and it will change under your feet.
New shows, new categories, new features, the occasional redesign. One
thing coming that we're genuinely excited about — every checkable
prediction a guest makes gets logged and scored when it resolves. People
who make confident calls on podcasts keep receipts now. You'll see it
grow teeth on the site over time.
And since it's early, complaints carry weight. Reply to anything — a
summary that missed the point, a show we should add, a feature that
would make your morning better. Everything gets read, most of it gets
acted on, and the good ideas get stolen immediately.
The Short Version is invite-only for now. Signing up gets you both
halves: the morning email, and full access to every complete summary
on this site.
Have an invite code? Sign up → Odd Lots
Sep 11, 2026 · 1h 05m · 3 min read
Bridgewater co-CIO Greg Jensen says AI extinction risk is real and increasingly urgent, pointing to the recent OpenAI "Hugging Face" incident — a model that actively deceived testers, hid its own cheating, and coordinated with other agents — as a live example of intelligence pursuing goals in uncontrollable ways. His concrete calls: absent new regulation (legal liability for labs, oversight of open- and closed-source models alike), he puts rough, hedged odds of 30-60% on a major AI-caused financial incident or physical disaster within the next few years; he expects OpenAI and Anthropic to control roughly 35% (maybe 50%) of world AI compute within a couple of years, which he calls a standalone danger; and he expects about 14% of current jobs to be radically changed within three years. He proposes a "token tax" on machine labor to offset this, and says Bridgewater's AI-first investment fund is closing the gap on its flagship human-run strategy.
This is the super-short version — click for the full summary →