The Beep That Keyed the Transmitters The short tone that cuts the recordings of Apollo voice loops is not a courtesy beep and not a signal that the speaker has finished. It is a pair of pure sine waves generated in Houston and sent down the same voice circuit that carried the CapCom’s words. One tone, 2,525 Hz for 250 milliseconds, told the remote ground station to key its high-power transmitter. The second, 2,475 Hz for the same duration, told the station to release the key. Both tones rode inside the voice band so NASA did not have to lease a second long-haul circuit simply to switch the transmitters on and off. CapCom sat in Houston and used a push-to-talk switch exactly as a pilot or a radio operator would. The switch did not connect directly to a transmitter. The active station might be in Australia, Spain, or California, chosen by the rotation of the Earth and the position of the spacecraft. When CapCom pressed the button, the tone generator in the Mission Control voice system inserted the higher-frequency burst. Detectors at the distant station recognized the tone, closed a relay, and placed the uplink audio on the S-band transmitter aimed at the spacecraft. When CapCom released the button, the lower-frequency burst arrived and the relay opened. The astronauts’ own transmissions needed no such tones; their microphones keyed local transmitters inside the spacecraft or on the lunar surface. The equipment that generated and detected the tones was built by Quindar Electronics. The name stuck. The same arrangement had already been used on Gemini and remained in service on early Shuttle UHF voice circuits. Full-duplex operation meant both ends could transmit at once on separate frequencies, so the tones did not mark the end of a transmission in the way a repeater’s courtesy tone does. They simply marked the state of CapCom’s switch. Notch filters at the uplink sites removed most of the tones from the audio that reached the spacecraft, which is why the crews rarely heard them even though the tones are prominent on the ground recordings that were later released. Frank Borman once noted that the intro tone could clip the first syllable if CapCom began speaking before the remote relay had closed. The fix was procedural: wait a fraction of a second after keying. The system was never elegant. It was the cheapest reliable way to control transmitters scattered across the globe when every additional leased line carried a real cost. The two beeps therefore did exactly the work a local microphone switch would have done had the transmitter been in the same building.

The recent XRP Ledger security fix raised a lot of questions and some misconceptions. Let's clear them up. 1. Was any XRP created? No. XRP's supply was fixed at 100 billion when the XRP Ledger launched, and no transaction is allowed to create more. This bug was a way around that rule, and 3.4.1 fixed it 2 weeks ago. You don't need to trust us on this. All ledger data is public, and explorers and data sites read it every day. New XRP would show up in account balances for everyone to see. There is no hidden place for XRP on the XRP Ledger, every balance is public. We monitor the XRP Ledger 24/7 and have verified the supply throughout the history of the ledger. Others in the community verified this too, and anyone can do the same. Transaction fees are destroyed, so the supply is deflationary. About 99.986 billion XRP exist today. 2. How did this go unnoticed, and what's next? No normal payment comes anywhere near triggering it. It took hundreds of deliberately constructed operations. AI makes bugs like this faster to find, for the bad guys and the good guys. Together with RippleX and the community, we will ensure the XRP Ledger is protected against such vulnerabilities. This means radically applying Formal Verification and AI adversarial testing, collaborating closely with security researchers, and strengthening the bug bounty program. You'll hear a lot more about this as we progress. 3. XRP Ledger decentralization at the core of this. A public fix shows attackers where the bug is, and AI makes that fast. With a normal amendment vote, the fix would have been public for weeks before it took effect, risking network security. As a reaction to this, the XRPL community came together, and among them the XRP Ledger Foundation, RippleX, core devs and validators formed a social consensus and agreed on a path forward that involves shipping the fix directly. Patching a live network is never trivial, so the plan was also designed to minimize disruption for everyone running on it. We shared the build with the default UNL validators, using the contact details each one publishes in its xrp-ledger.toml file. Public binaries followed the same day, September 25. No single party could make that call alone. Every validator and node operator chooses what software to run, and 3.4.1 only became required through a normal amendment vote with more than 80% validator support for two weeks. Protocol changes still go through amendments. Big thanks to Cayden Liao and Veria AI for finding and reporting this responsibly, to the XRPL core developers for building and testing the fix, and to every node operator who upgraded so quickly, especially the validators. Full report:

The real risk in AI agentic commerce isn't the stolen $320M that made headlines last month. The blockchain money is traceable and probably coming back. But the 67,000 Trezor hardware wallet customers whose home addresses leaked through a shipping vendor stay public forever. “A stolen coin can be returned, a leaked identity cannot,” says our CEO Evin McMullen on @CoinDesk. A leak is worse than a hack. Now we're about to hand the same data-hoarding architecture to billions of AI agents. The fix already exists, we're just not using it.

The Fix Uxbridge Road campaign argues that the stretch between Shepherd’s Bush and Acton Vale is one of the 'hollowed out' high streets that Andy Burnham targeted when he announced a crackdown on vape stores, betting shops and 'rogue businesses'.

The bottleneck in most AI-heavy businesses is approving much more so than generating. Just delivered an assessment to a founder who already runs Claude Code, a second brain on a Mac Mini, and a one dedicated wiki per client. His constraint is that every call summary and framework has to be manually reviewed before it enters the knowledge base. Generation is instant but ratification piles up. The fix, in two sizes: Quick win (this week) -Add his assistant to TailScale so she can reach the Mac Mini -She runs the existing intake skill on each Fathom transcript -Output goes to him for a yes or no, then into the second brain Major project (later) -Fathom fires a webhook when a transcript is done -An agent listens and runs the intake skill automatically -A routing skill sends the result to the right wiki (main brain first, then the client's) -Assistant approves or rejects in one shared channel Net savings we estimated: 1.5 to 2 hours a week after 30 to 60 minutes of review time.

The reason the years start flying by is because you stop doing anything new. Think about why childhood felt so long, everything was a first with new places, feelings and experiences. Your brain was recording all of it because none of it was familiar yet. The older you get, settle into a routine, every week starts looking the same. Wake up, phone, work, scroll, sleep, repeat. Your brain stops giving a shit to record it because there's nothing worth remembering anymore. That's why a month can vanish and you can't recall a single day of it. The fix isn't some big life change btw, it's the super simple things. Take a different route home, go somewhere you've never been, meet new people, try something you're bad at and put yourself in unfamiliar situations. Anything that snaps you out of autopilot. The more new experiences you pack in the slower and fuller time feels. Go do something you'll actually remember.

【#THEFIRSTTAKE】# ー No.688 TAEYANG / LIVE FAST DIE SLOW ー A microphone and a white studio. And 1 rule. You’ve got 1 TAKE. ・ @Realtaeyang ▼Full show on YouTube

The fix: Download BleachBit. (Open-source). It scans for caches, temporary files, logs, and leftover data from dozens of programs. Select what you want gone. Preview the file list. Hit Clean.