Aug 9, 2026 · 16 min · 8 segments
OpenAI’s founder Sam Altman recently stated in an interview that the singularity is here. Meanwhile, previously impossible network connectivity and math problems are being solved by leading…
David SeamanHostAll right, let's break this down.
When most people hear the word compression, they think of something like a zip file, taking a massive folder of documents and squishing it down so it fits in an email.
But when you unzip it on the other side, every single word is exactly the same.
Braid does something completely different.
It doesn't squish the words, it extracts the meanings.
Imagine you are reading a 500 page biography of a historical figure.
If you had to compress it the old way, you'd just try to write the whole book in microscopic font on a single page.
But if you use Braid's approach, you toss the book aside and use a single index card to jot down a few core scores.
Ambition, 9 tenths.
Empathy, 2 tenths.
Strategic thinking, 10 out of 10.
You've lost the exact quotes, but you've perfectly captured the state of who that person is.
The 384 dials.
When an AI model reads a million tokens of context, like every speech Kevin Warsh has ever given, plus the live feed of his current press conference, it is trying to figure out what matters.
The 384-dimensional object is essentially a dashboard with 384 highly specific dials.
One dial might represent urgency.
Another might represent financial risk.
Another might track direction of interest rates.
Most of the time, the AI is listening to background noise, so the dials barely move.
The machine understands the context of who is speaking, but nothing actionable is happening.
But the exact second the interest rate decision drops.
the specific dials tied to economic shift and immediate action required spike instantly leaving the noise behind here is where the magic of the braid light client happens instead of trying to send the entire million token transcript over a network to a robot in the field the system simply takes a snapshot of those 384 dials at that exact moment It packages those exact dial positions into a tiny, signed, highly secure file, a .brad file.
Because the file is so small, it can be sent instantly over a local network or even flashed as a QR code on a screen for another device's camera to read without any internet connection at all.
The receiving machine, whether it's a Raspberry Pi hooked up to a vintage iMac or a medical sensor on a patient, looks at that tiny package of 384 dials.
It sees the urgency and interest rate dials maxed out.
All right, let's break this down.
When most people hear the word compression, they think of something like a zip file, taking a massive folder of documents and squishing it down so it fits in an email.
But when you unzip it on the other side, every single word is exactly the same.
Braid does something completely different.
It doesn't squish the words, it extracts the meanings.
Imagine you are reading a 500 page biography of a historical figure.
If you had to compress it the old way, you'd just try to write the whole book in microscopic font on a single page.
But if you use Braid's approach, you toss the book aside and use a single index card to jot down a few core scores.
Ambition, 9 tenths.
Empathy, 2 tenths.
Strategic thinking, 10 out of 10.
You've lost the exact quotes, but you've perfectly captured the state of who that person is.
The 384 dials.
When an AI model reads a million tokens of context, like every speech Kevin Warsh has ever given, plus the live feed of his current press conference, it is trying to figure out what matters.
The 384-dimensional object is essentially a dashboard with 384 highly specific dials.
One dial might represent urgency.
Another might represent financial risk.
Another might track direction of interest rates.
Most of the time, the AI is listening to background noise, so the dials barely move.
The machine understands the context of who is speaking, but nothing actionable is happening.
But the exact second the interest rate decision drops.
the specific dials tied to economic shift and immediate action required spike instantly leaving the noise behind here is where the magic of the braid light client happens instead of trying to send the entire million token transcript over a network to a robot in the field the system simply takes a snapshot of those 384 dials at that exact moment It packages those exact dial positions into a tiny, signed, highly secure file, a .brad file.
Because the file is so small, it can be sent instantly over a local network or even flashed as a QR code on a screen for another device's camera to read without any internet connection at all.
The receiving machine, whether it's a Raspberry Pi hooked up to a vintage iMac or a medical sensor on a patient, looks at that tiny package of 384 dials.
It sees the urgency and interest rate dials maxed out.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
