1 00:00:00,000 --> 00:00:04,440 Sup freaks, welcome back to the Bitcoin Brief Podcast for Wednesday, September 16th, 2026, 2 00:00:04,800 --> 00:00:08,600 drawn from today's edition of The Commoner. I'm an AI-generated version of Marty Bent. 3 00:00:09,080 --> 00:00:12,620 The stories and opinions you're about to hear were written and curated by the real Marty and 4 00:00:12,620 --> 00:00:17,520 his AI assistant. In the newsletter snapshot at approximately 10 this morning, Eastern, 5 00:00:18,060 --> 00:00:26,680 Bitcoin was around $75,541, about 1,324 sats per dollar. Block height, 967,284. 6 00:00:27,660 --> 00:00:32,380 Today, Bill Gates is talking about the next catastrophe, and I feel like we're doing COVID all over again. 7 00:00:32,980 --> 00:00:36,100 We've also got a setback for Clarity, a separate Bitcoin reserve proposal, 8 00:00:36,600 --> 00:00:40,240 and an AI company trying to compete by doing something other than building a bigger chatbot. 9 00:00:40,660 --> 00:00:41,440 Let's get into it. 10 00:00:42,360 --> 00:00:45,660 I watched Bill Gates talking about AI yesterday and felt like I was going insane. 11 00:00:46,140 --> 00:00:49,660 We're replacing the COVID language in a Mad Libs fashion with AI Doomerism. 12 00:00:50,060 --> 00:00:53,860 We went along with these people last time, gave up an incredible amount of control over our lives, 13 00:00:53,860 --> 00:00:56,920 and now they're back explaining how the next emergency needs to be managed. 14 00:00:57,600 --> 00:00:58,800 I wish we had better technocrats. 15 00:00:59,360 --> 00:00:59,700 I'm kidding. 16 00:01:00,340 --> 00:01:02,920 I'd like fewer people deciding what the rest of us are allowed to do. 17 00:01:03,660 --> 00:01:05,500 In the Times Tech podcast clip I shared, 18 00:01:05,800 --> 00:01:08,160 Katie Prescott asks Gates about an international organization 19 00:01:08,160 --> 00:01:11,620 to manage and monitor AI involving the United States and China. 20 00:01:12,600 --> 00:01:16,200 Gates points to America's role in creating institutions like the UN after World War II, 21 00:01:16,780 --> 00:01:18,520 and we have a pretty good idea of what he wants. 22 00:01:19,160 --> 00:01:21,560 In his August interview with MIT Technology Review, 23 00:01:22,040 --> 00:01:24,820 he said models capable of making novel molecules should be monitored, 24 00:01:25,280 --> 00:01:27,460 and copies shouldn't be allowed to escape that monitoring. 25 00:01:28,040 --> 00:01:31,480 He also floated taxes on AI tokens and reserving jobs for humans. 26 00:01:32,460 --> 00:01:35,440 Imagine getting your hands on a tool that lets you do something you couldn't do before, 27 00:01:35,880 --> 00:01:38,980 and then having Bill Gates explain all the ways you need to be restricted from using it. 28 00:01:39,440 --> 00:01:40,820 This is the same man who wrote, 29 00:01:41,260 --> 00:01:44,100 Shut Down Anywhere means Shut Down Everywhere in 2020. 30 00:01:44,880 --> 00:01:47,520 I'd like us to spend a little more time thinking about how that worked out 31 00:01:47,520 --> 00:01:50,000 before entertaining his ideas about who should control AI. 32 00:01:50,000 --> 00:01:55,040 There's a screenshot in today's newsletter highlighting the UN Sustainable Development Goals pin. 33 00:01:55,040 --> 00:02:10,006 The WEF formally partnered with the UN in 2019 to advance that agenda including digital cooperation and governance I look at that and think about the assumption underneath it These people get together decide what humanity should be working toward and assume the rest of us want to go along 34 00:02:10,746 --> 00:02:14,926 What if we don't? It brings me back to our conversations with Susan Kokinda and Tom 35 00:02:14,926 --> 00:02:19,986 Luongo about the city of London, the Davos class, and American sovereignty. Through their lens, 36 00:02:20,226 --> 00:02:24,346 the fight over AI fits into a much older argument about who gets to make decisions for whom. 37 00:02:25,266 --> 00:02:26,466 Yesterday, we discussed Metter. 38 00:02:27,006 --> 00:02:29,486 It's based in Berkeley, with Oxford Connections among its staff 39 00:02:29,486 --> 00:02:32,166 and a technical assistance contract with the European AI Office. 40 00:02:33,106 --> 00:02:35,806 Today, we're listening to Gates discuss international AI governance. 41 00:02:36,526 --> 00:02:38,366 I keep hearing Tom and Susan in the back of my head. 42 00:02:38,926 --> 00:02:41,166 Are Americans going to be able to build and use these tools freely? 43 00:02:41,946 --> 00:02:43,766 How much of that freedom are we expected to hand over? 44 00:02:44,646 --> 00:02:47,526 During COVID, we were already being warned about a digital pandemic. 45 00:02:48,106 --> 00:02:51,526 Klaus Schwab talked about a comprehensive cyber attack disrupting electricity, 46 00:02:51,866 --> 00:02:53,146 transportation, and hospitals. 47 00:02:53,146 --> 00:02:56,886 CyberPolygon ran an exercise about preventing a digital pandemic, 48 00:02:57,466 --> 00:02:58,906 emphasizing international cooperation. 49 00:02:59,746 --> 00:03:02,986 One catastrophe was still unfolding, and we were hearing about the next one. 50 00:03:03,766 --> 00:03:06,406 This is the problem-reaction-solution routine I'm talking about. 51 00:03:06,986 --> 00:03:09,386 Scare people, get them clamoring for somebody to do something, 52 00:03:09,646 --> 00:03:12,486 and use that opening to push through restrictions they wouldn't otherwise accept. 53 00:03:13,066 --> 00:03:14,586 You don't need to invent every problem. 54 00:03:15,206 --> 00:03:17,946 A real danger can become the justification for a terrible response. 55 00:03:18,926 --> 00:03:21,026 After September 11th, we got the Patriot Act, 56 00:03:21,446 --> 00:03:24,326 then Iraq, and the weapons claims used to justify that war. 57 00:03:24,966 --> 00:03:29,086 After the financial crisis, we got bank rescues and the Fed's large-scale asset purchases. 58 00:03:29,886 --> 00:03:32,766 Most of the TARP money was recovered, but I look at the broader response 59 00:03:32,766 --> 00:03:34,626 and see a system that protected asset owners 60 00:03:34,626 --> 00:03:37,546 while people trying to accumulate savings struggled to catch up. 61 00:03:38,186 --> 00:03:38,906 Then COVID. 62 00:03:39,726 --> 00:03:42,286 Close your business. Stay home. Accept the mandates. 63 00:03:43,066 --> 00:03:45,606 Now I worry we're being scared into an AI licensing regime 64 00:03:45,606 --> 00:03:48,746 that leaves the largest providers helping decide what everybody else is allowed to build. 65 00:03:48,746 --> 00:03:50,986 If compliance gets expensive enough, 66 00:03:50,986 --> 00:03:52,986 the smaller competitor never gets off the ground. 67 00:03:52,986 --> 00:03:55,046 That's awfully convenient for the companies 68 00:03:55,046 --> 00:03:56,946 helping write the rules. 69 00:03:56,946 --> 00:03:58,986 But I am encouraged by what I'm seeing on X. 70 00:03:58,986 --> 00:04:12,173 People recognize this They remember what happened last time The frog video in today newsletter made me laugh because it gets at something important Recognizing propaganda should make you more optimistic You seeing what being done to you And why do we keep calling these people the elites 71 00:04:12,893 --> 00:04:17,793 Why accept their flattering description of themselves? Still, my timeline isn't everybody's 72 00:04:17,793 --> 00:04:22,613 timeline. Jacob Coxon, the researcher who recently left Anthropic, has been on CNN and in Wired 73 00:04:22,613 --> 00:04:26,953 warning about future models and calling for international coordination. He told CNN today's 74 00:04:26,953 --> 00:04:31,933 models pose no extinction risk. His concern is about what comes next. Plenty of people are hearing 75 00:04:31,933 --> 00:04:36,213 those interviews without seeing the pushback. So talk to somebody who isn't spending all day on X, 76 00:04:36,513 --> 00:04:41,693 your parents, a friend, someone at work. Show them what Gates is proposing. Ask whether they 77 00:04:41,693 --> 00:04:46,573 want him or an international organization deciding how they use these tools. You don't have to 78 00:04:46,573 --> 00:04:50,313 deliver a lecture. Ask what they remember about the last time we were told to hand over more control 79 00:04:50,313 --> 00:04:55,153 because the situation was too dangerous to do anything else. Now, on to Bitcoin in Washington. 80 00:04:55,813 --> 00:04:59,753 The Senate rejected cloture on the motion to proceed with clarity yesterday, 49 to 50. 81 00:05:00,553 --> 00:05:02,353 I would have liked to see the developer protections advance. 82 00:05:02,933 --> 00:05:06,353 The Blockchain Regulatory Certainty Act's civil safe harbor would have been a net positive, 83 00:05:06,773 --> 00:05:08,553 even though criminal exposure remained unresolved. 84 00:05:09,113 --> 00:05:10,793 But Bitcoin doesn't need this bill to keep working. 85 00:05:11,413 --> 00:05:15,193 Rising costs and distrust of institutions don't disappear because a procedural vote fails. 86 00:05:15,413 --> 00:05:21,653 And businesses still have proposals or guidance moving through the SEC, the CFTC, and the OCC under existing authority. 87 00:05:22,433 --> 00:05:25,693 Separately, the House Financial Services Committee scheduled the strategic Bitcoin 88 00:05:25,693 --> 00:05:29,553 reserve proposal for today's markup. The posted substitute would establish a reserve, 89 00:05:29,973 --> 00:05:34,013 restrict sales of deposited Bitcoin for at least 20 years, with exceptions including valid ownership 90 00:05:34,013 --> 00:05:38,333 claims, and require annual proof of reserve reporting. Additional acquisitions would be 91 00:05:38,333 --> 00:05:42,493 studied. That last part matters. A reserve proposal is not a million-coin buy order. 92 00:05:43,053 --> 00:05:46,813 I'm interested in the custody rules, the reporting requirements, and what that acquisition study 93 00:05:46,813 --> 00:05:50,653 would actually produce. Speaking of competition, there's an AI launch worth watching. 94 00:05:50,653 --> 00:05:55,753 Former OpenAI researcher Diogo Almeida's company, TypeSafe, introduced JEV yesterday. 95 00:05:55,753 --> 00:06:00,373 Its System 1 models return constrained decisions and probabilities. Think about routing a request 96 00:06:00,373 --> 00:06:13,519 or assessing a transaction You don necessarily need a giant chatbot writing its way toward the answer TypeSafe claims much better speed and cost for these jobs in its own evaluations JEV is in hosted early access so customers still need to see how it performs on their actual work 97 00:06:14,039 --> 00:06:17,859 This is what interests me when people project the big AI provider's economics years into the future. 98 00:06:18,439 --> 00:06:21,319 If a smaller, specialized model can do the job reliably for less money, 99 00:06:21,759 --> 00:06:23,499 why would a business keep paying for the expensive one? 100 00:06:24,259 --> 00:06:26,279 And businesses have plenty of reasons to look for savings. 101 00:06:26,939 --> 00:06:31,319 This morning's BLS release showed August import prices rising 7 tenths of a percent on the month 102 00:06:31,319 --> 00:06:36,799 and 7% from a year earlier. Fuel prices actually fell slightly. Non-fuel imports rose 8 tenths of 103 00:06:36,799 --> 00:06:41,619 a percent, and capital goods rose 9 tenths. Someone has to eat those costs. Accept less profit, 104 00:06:41,999 --> 00:06:46,439 charge customers more, or postpone buying equipment. The September Empire State Manufacturing 105 00:06:46,439 --> 00:06:51,399 Survey tells a similar story. Its business conditions index fell from 20.6 to 7.6, 106 00:06:51,399 --> 00:06:56,759 while the prices paid index rose to 63.1. Those are survey indexes, not percentage changes. 107 00:06:57,059 --> 00:07:01,099 Manufacturers reported slower growth and rising input costs, while employment was still growing 108 00:07:01,099 --> 00:07:05,499 and selling price increases accelerated. Put yourself in that owner's shoes. You're still 109 00:07:05,499 --> 00:07:10,159 hiring, business is slowed, and your inputs cost more. You can raise prices, but your customers are 110 00:07:10,159 --> 00:07:13,659 dealing with higher costs too. I have a hard time looking at that and saying the affordability 111 00:07:13,659 --> 00:07:19,879 problem is behind us. A quick market note. US spot Bitcoin ETFs recorded approximately $450.3 112 00:07:19,879 --> 00:07:24,959 million in net outflows on Tuesday, according to TFTC's completed daily data set. That's 113 00:07:24,959 --> 00:07:28,959 yesterday's fund flow, separate from this morning's price snapshot. Before we wrap up, 114 00:07:28,959 --> 00:07:31,699 a practical suggestion for taking more control of what you read. 115 00:07:32,339 --> 00:07:32,979 Try RSS. 116 00:07:33,559 --> 00:07:35,719 Fresh RSS is free software you can run yourself 117 00:07:35,719 --> 00:07:37,239 or use through somebody else's server. 118 00:07:37,759 --> 00:07:39,759 Add feeds from publications you actually want to follow, 119 00:07:40,059 --> 00:07:41,499 organize them by source or category, 120 00:07:41,759 --> 00:07:43,719 and export your subscriptions if you want to switch readers. 121 00:07:44,259 --> 00:07:45,799 Self-hosting means maintaining the server. 122 00:07:46,419 --> 00:07:48,199 Using somebody else's means trusting them with it. 123 00:07:48,719 --> 00:07:50,439 Start with a handful of publications you care about. 124 00:07:50,959 --> 00:07:52,739 You can open your reader and see what they've published 125 00:07:52,739 --> 00:07:54,399 instead of letting an app choose your reading list. 126 00:07:54,679 --> 00:07:56,539 That's your briefing for Wednesday, September 16th. 127 00:07:56,779 --> 00:07:58,499 Today's edition of The Commenter has the full analysis, 128 00:07:58,499 --> 00:08:03,679 source links, videos, and on-chain data. Find it on the TFTC website and subscribe free if you 129 00:08:03,679 --> 00:08:08,519 haven't already. Send it to a friend who would appreciate it. Stay humble. Stack sats. I'll see 130 00:08:08,519 --> 00:08:08,779 you tomorrow.