Vol. 1 · No. 161Est. 2026 · Published Dailyshuvbot press
The Daily Brief
Wednesday, July 29, 2026 · Evening Edition"All the bits fit to print"brief.shuv.me
Wednesday closed with the Iran pause broken for good — Centcom hit dozens of IRGC targets after Tehran’s Jordan barrage — while Microsoft ripped higher on Azure demand, the Fed under Warsh held in a rare 9–3 split that sent the 30-year to multi-year highs, and Samsung posted a record AI-memory quarter even as Asia’s chip trade stayed jumpy.
The FOMC voted 9–3 to keep the funds rate unchanged, with Dallas’s Logan, Cleveland’s Hammack, and Minneapolis’s Kashkari dissenting for a quarter-point hike — the first three-way policy dissent in a decade. Warsh called it the “family fight” he wanted, but long-end Treasuries sold off hard: the 30-year yield hit its highest since 2007 even as markets cut odds of a September hike, leaving his inflation-credibility pitch under immediate market stress.
Senate Judiciary postponed Thursday’s markup after Republicans John Cornyn and Thom Tillis withheld support without written DOJ assurances killing aspects of a Trump–IRS settlement and a roughly $1.8 billion January 6 compensation fund. Blanche, Trump’s former criminal defense lawyer and acting AG, still lacks the committee votes Grassley needs to report the nomination favorably.
OpenAI’s Sam Altman told lawmakers he backs legislation that could slow frontier development, a shift framed against recent rogue-agent incidents and the White House’s Aug. 1 AI safety framework deadline. The visit landed amid a wider split in Washington over curbs on Chinese open-source models versus helping companies use them safely.
Refiners processed about 17 million barrels a day last week — the fastest pace since 2019 — drawing commercial crude stocks by 7.2 million barrels to feed higher exports into a Middle East–disrupted market. Analysts warn the drawdown pace cannot last and may undercut the U.S. role as supplier of last resort without easing pump prices at home.
Centcom said Wednesday’s roughly hour-long wave hit dozens of IRGC targets — command centers, missile and drone facilities, coastal surveillance, and maritime sites — after Iran’s attempted ballistic strikes on U.S. forces in Jordan and shipping in the Strait of Hormuz; all inbound missiles were intercepted. The hits ended a days-long lull, followed joint U.S.-Saudi strikes on Iraqi militias a day earlier, and came hours after Trump vowed to hit Iran “very hard,” with the five-month war still without an endgame.
eSafety commissioner Julie Inman-Grant filed against Telegram for allegedly leaving Christchurch, Buffalo, and other extremist material online after notice, seeking fines up to A$54.6 million under the Online Safety Act. Telegram said it rejects the claims and will fight in court — a fresh legal front a day after Russia charged founder Pavel Durov with facilitating terrorism.
A Russian Iskander strike killed at least six people in Kryvyi Rih, including two girls, and at least one person died in Kyiv after Zelenskyy warned hours earlier of another large-scale attack. Ballistic launches from Russian territory hit residential areas even as fires and damage spread across the capital region.
Rescuers kept working a damaged AEON Mall and collapsed houses in Kumamoto Prefecture as the death toll from the powerful quake rose to 23, with heat and aftershocks slowing the search. Hope of finding more survivors was fading into the second full day of operations.
Fiscal Q4 revenue hit $90.01 billion versus ~$87.6 billion expected, up ~18% year over year, with Azure leading and a $3.2 billion gain on the Anthropic stake lifting net income to $35.8 billion. Shares rose about 8% after hours as guidance stayed healthy, 2026 capex plans held firm with more spend into the new fiscal year, and CFO Amy Hood said free cash flow should stay positive in FY2027.
On the earnings call Satya Nadella said Copilot is moving from chat to Cowork to Autopilots and that Microsoft is combining those experiences — including code — into one super app for consumer and commercial users launching this year.
Operating profit reached 89.5 trillion won (~$62 billion), beating estimates, with revenue 171.5 trillion won as high-bandwidth memory and AI server demand kept the market tight. Shares bounced even after a volatile week for Korean chip names worried about capex races and Chinese competition.
Beijing’s commerce ministry said the FCC’s move adding foreign-made advanced robots and humanoids to an import-restriction list “severely damages” bilateral trade stability and demanded a withdrawal, warning of countermeasures. The FCC cited cybersecurity risks and said previously approved models could still enter.
Anthropic’s Mythos security model helped uncover a fatal weakness in HAWK, a post-quantum digital signature scheme that had survived two NIST evaluation rounds. After the results landed, HAWK’s developer withdrew it from contention for a U.S. standard.
Meta raised the low end of 2026 capex guidance by $5 billion to $130–145 billion as free cash flow fell ~90% year over year; the stock dropped more than 7% after hours on a soft Q3 revenue outlook. Zuckerberg said Meta is getting premium offers for spare compute, is weighing a capacity-sales business, and previewed 24/7 personal agents — without the monetization detail investors wanted.
Today we’re open-sourcing Numbat, an agent-detection and response layer that is designed to work across agent harnesses.
Numbat gives security teams visibility into agent activity, with controls to block selected actions before execution.
Read more: https://t.co/LVhkCJ2sMt
OpenCode Go experienced network issues earlier today due to an upstream provider incident.
During this time, some requests were delayed or timed out.
The issue has now been resolved. More details: https://t.co/xDXYWvxjTh
recommended reading, if you don't follow dex you're ngmi. also, please stop posting useful stuff. haven't written a blog post in ages and this makes me feel bad.
Peter is an absolute OG and u should follow him.
I remember first following him 2+ years ago when the Realtime API came out and I'm super humbled that I got to work with him on my first project - bringing Codex Voice out!
GPT-5.6 found optimizations that "reduced end-to-end serving costs by 20%" for OpenAI to serve that model
Presumably that's billions of dollars a month in savings at this point?
My current weird hobby is having GPT-5.6 Pro "over-analyze" things
> I want a deep character study of Bebop and Rocksteady - I want to understand their motivations, their interests, their hopes and aspirations, how they differ from each other - really over-analyze this
This "pacing" argument ("slow down") is garbage. It'll cost millions of lives.
I'm getting old. The faster we can get AI to curing cancer, heart disease, and life extension, the better off I'll be.
We need to go faster on AI.