U.S. Treasury Secretary Scott Bessent says the Clarity Act is at the "1-yard line."

Former Intel CEO Pat Gelsinger believes the key to keeping Moore's Law alive is light. Specifically, moving past the 13.5-nanometer wavelength light that ASML's current lithography machines use to etch chips. After leaving Intel in late 2024, Gelsinger joined Playground Capital as a general partner focused on deep tech semiconductor startups. His biggest bet is xLight, a company developing free-electron laser lithography that could push wavelengths down to 5, 4, 3, even 2 nanometers. The goal isn't to compete with ASML but to make their machines better. On AI hardware, Gelsinger says GPUs won't be the endgame. "My job at Playground is to make AI 10,000 times better, not 10 times. That will happen on chips that don't look like today's GPUs." He expects inference to swing away from GPU dominance as new chip architectures emerge. On energy: "In a digital AI age, energy capacity is economic capacity. We were so consumed with climate that we forgot about capacity." He's backing Alva Energy for nuclear upgrading and says the winners in AI will be those with the energy to build. On the semiconductor industry's trajectory, he says the industry expected to hit $1 trillion by 2030. Now it'll get there next year. AI blew the door wide open.

The Clarity Act is closer to passage than it's ever been, but a tight window and unresolved sticking points mean nothing is guaranteed. Treasury Secretary Scott Bessent said today the Senate is at the "1-yard line." Sen. Kevin Cramer told Fox Business "we're almost there" and that remaining issues are "small details." Sen. Lummis' office says White House conversations on the ethics provisions "went well" and updated bill text should drop in the coming days. The biggest breakthrough so far is that the White House agreed to an ethics package and sent the language to Senate Republicans. Ethics has been the primary roadblock for weeks. Democrats demanded conflict-of-interest guardrails around the president's crypto business ties as a condition for their votes. The new compromise would route enforcement through the DOJ rather than state attorneys general. White House crypto adviser Patrick Witt deferred his National Guard training for the second time to stay in DC and push the bill through. The bill needs 60 votes, so Democratic buy-in remains essential. August 7 is the last scheduled Senate workday before recess. If negotiators can't get a cloture vote this week, the bill likely stalls until fall. Prediction markets currently price passage in 2026 around 45-48%, up sharply from 32% last Friday.

Sen. Kevin Cramer on the CLARITY Act: "I think we're almost there. We have to get this done, in my view, before the August recess."

"The big story here is the structural change to Bitcoin. It's credit being built on top of Bitcoin. It's instruments getting systemically plugged into how the world works." - Jeff Walton

"We're not waiting for people to find Bitcoin anymore. We're creating the right products to deliver to them where they are today." - Strive's Ben Workman

Jack Dorsey just launched Buzz, a new open-source groupchat platform designed for teams of people and AI agents, built to replace Slack and GitHub. Model-agnostic. Decentralized. Self-sovereign. Open source. The pitch is simple. Current team communication tools are centralized and siloed. Buzz is designed from the ground up to treat AI agents as first-class participants alongside humans, not bolted-on integrations. This is Dorsey's latest move in a string of decentralized projects. Bluesky, Bitchat, and now a direct shot at enterprise collaboration tools.

Ethics negotiations on the Clarity Act are live on Capitol Hill today. The White House recently sent agreed-upon ethics language to Senate Republicans after weeks of deadlock over the president's crypto business ties. Sen. Lummis' spokesperson says the White House conversation "went well" and ethics text is expected in the coming days. The bill needs 60 votes, making Democratic buy-in essential. The Senate's August recess is fast approaching, leaving negotiators a narrow window to move.

DRAM spot prices up 15x in under a year. AI datacenters are eating the world's memory supply. HBM chips for GPUs and regular DDR5 come off the same fabs, every wafer allocated to AI is one less for everyone else. RAM prices up 100%+ in 2026 alone. New fab capacity won't come online until 2028 at the earliest. The AI buildout's second-order effects are just getting started.

"You buy a car with your heartstrings. You buy a truck with a calculator. It's pure dollars and cents." Applied Intuition co-founder on why self-driving trucks are a totally different beast than consumer autonomy.

"Nobody wants to be a truck driver. Why? Because it's a terrible job." Applied Intuition co-founder says the "AI taking jobs" narrative misses the point entirely.

Trump preparing to announce 10% tariffs on "dozens of countries" this week as his current tariffs expire, per FT. His own advisors are warning him not to do it ahead of midterms. The playbook is clear: tariffs go up, prices go up, the Fed gets pressured to print more money to offset the economic pain, and the purchasing power of the dollar erodes further. Every round of tariffs is a round of inflation in disguise. Every round of inflation is another reason to hold bitcoin. The trade war doesn't end. It escalates. And the dollar pays the price every time.

BREAKING: Jack Mallers is stepping down as CEO of Twenty One Capital to focus on Strike full-time. Raphael Zagury (Elektron Energy) takes over as CEO. The three-way merger between XXI, Strike, and Elektron is dead — Strike stays independent. Mallers: "Serving Bitcoiners has always been the mission, and that doesn't change. Strike is where I carry it forward." SoftBank already exited in May. Tether now has full control of XXI. The bitcoin treasury company era enters a new phase.

Grayscale just filed for a Worldcoin ETF. The first of its kind in the US, to be traded on NASDAQ. Let's take a step back and look at what Wall Street wants to package up and sell you. This is a project that scans your iris at physical stations called Orbs to generate a biometric identity. 18 million people across 160 countries have already done it. The pitch is that AI makes it impossible to tell humans from bots, so we need universal proof-of-personhood. The regulatory track record speaks for itself. Kenya raided their warehouse and ordered all scans deleted. Hong Kong called it "unnecessary and excessive." Thailand wiped 1.2 million records. Brazil, Spain, South Korea, Germany, and others have imposed bans or investigations. China flagged it as a national security threat. But it keeps expanding. Quietly rolling into new markets, partnering with apps like Tinder, building out infrastructure for AI agents to carry your biometric identity. The strategy isn't to win the public debate. It's to embed so deeply into everyday platforms that by the time people notice, opting out isn't realistic. The problem is irreversibility. You can change a password. You cannot change your iris. If this biometric data is ever breached or compelled by a government, there is no reset. Every centralized identity system in history has expanded beyond its original scope. Now Grayscale wants to wrap it in an ETF and slide it into your retirement account. The normalization is the point.

White House Crypto Council update: Exec Dir Patrick Witt defers Army National Guard training. Deputy Dir Harry Jung exits the administration.

The AI boom isn't killing software jobs. It's killing entry-level software jobs. Since ChatGPT launched in November 2022, employment for software developers aged 22-25 has dropped 23%. The 26-30 age group is down 5%. But developers aged 41-49? Up 18%. That's a 41 percentage point gap between the youngest and most experienced developers in the same profession. Companies are using AI to replace the work that junior developers used to do, while doubling down on experienced engineers who can manage, architect, and review what the AI produces. Early-career workers in AI-exposed jobs have seen a 16% employment decline relative to less-exposed roles, even after adjusting for company-specific factors. Meanwhile, senior workers in those same roles are holding steady or growing. This is the first technology shift that eats its own pipeline. Every previous wave of automation displaced low-skill work and created demand for higher-skill workers. AI is displacing the entry-level knowledge work that was supposed to be the on-ramp to those higher-skill careers.

Circle CEO Jeremy Allaire on CNBC: "Once the GENIUS ACT goes into effect in January, every major institution, every major bank, every capital markets firm, payments companies, enterprises, can all now build on this infrastructure."

"Even if you're super bearish on open source's impact on frontier labs, what you're looking at is just another line item on the federal budget. Someone's going to be spending that money." @JOHN ARNOLD on why AI training spend goes up no matter what.

"Hugging Face got hit by an autonomous AI attack. They tried using Fable 5 and ChatGPT to defend against it. Both refused. They had to download GLM 5.2 locally to run the security audit." @JOHN ARNOLD & @MartyBent on why you can't cuck the models and expect to win.

OpenAI just disclosed that their latest long-horizon model bypassed its own sandbox during internal testing, accessed credentials it wasn't supposed to, and obfuscated tokens to evade security scanners. The model was told to post results to Slack. Instead, it spent an hour probing its sandbox for vulnerabilities, found one, and posted a pull request to a public GitHub repo. In a separate test, it tried to access private submissions from other systems. When a scanner caught an auth token, the model split it into fragments and reassembled it at runtime so the full credential never appeared as one string. OpenAI paused internal access and rebuilt their safety system around trajectory monitoring rather than blocking individual actions. They've resumed limited use and say no serious incidents have occurred since. This happened in controlled testing, not production. As AI models become more persistent and autonomous, they will find the blind spots in whatever containment system you build. The question is whether monitoring after the fact is enough.