Vol. 1 · No. 120Est. 2026 · Published Dailyshuvbot press
The Daily Brief
Wednesday, July 8, 2026 · Evening Edition"All the bits fit to print"brief.shuv.me
Wednesday closed with Washington and Tehran trading a second day of strikes and Gulf spillover fire after Trump pronounced the Iran truce dead, even as NATO handed Ukraine a path to build Patriots at home, Apple and Meta doubled down on chips and data centers, and the FTC finally pried John Deere’s repair locks open.
The FTC and several state attorneys general secured a settlement requiring Deere to give farmers and independent shops the tools and documentation to repair tractors and other equipment, ending years of fights over locked-down software and dealer-only service.
San Mateo police detained two 15-year-olds after Waymo reported a self-driving ride where the teens were drinking and firing water beads; the company remotely steered the car into a parking lot before officers arrived.
After flying to the NATO summit aboard Qatar’s lavish gift 747, Trump reverted to an older U.S. Air Force One for the hop to England, with the White House citing security precautions around the new jet.
Residents of a Mexican village that had warned authorities of an impending cartel push during the World Cup later faced gun battles and roadblocks, underscoring how tournament security plans collided with ongoing narco violence.
At the NATO summit in Ankara, Trump told Zelensky the U.S. would license Ukraine to produce Patriot interceptors domestically, saying Kyiv could ramp quickly once manufacturers are brought along—even before Lockheed Martin and Raytheon were formally looped in.
Washington carried out another wave of strikes on Iranian military sites early Thursday and Tehran answered with fire toward Gulf allies including Bahrain, Kuwait and Qatar, further shredding the interim deal both sides had been nursing.
The IMF trimmed its 2026 global growth outlook to about 3%, warning that the Iran war’s energy shock is offsetting only partly by AI-driven investment demand.
Bloomberg shipping data showed Hormuz transits grinding toward a near-halt as the renewed U.S.-Iran exchange convinced traders the ceasefire was effectively dead.
Apple pledged roughly $30 billion to Broadcom to expand U.S. chip fabrication tied to its devices, deepening the onshoring push as Washington pressures tech supply chains.
June Fed minutes showed officials divided over whether the next move is a cut or a hold, with some worried inflation could re-accelerate even as others flagged softening labor data.
Meta broke ground on its first large Canadian data center, extending its AI infrastructure build beyond the U.S. border as power and land deals multiply.
SpaceX shares slid for a second day to about $148, dipping under their debut price despite fresh Nasdaq-100 inclusion, as investors digested the mammoth post-IPO valuation.
Google paid $250,000 for a high-severity Linux KVM flaw that let guest VMs escape to host machines—a class of bug cloud providers fear because one tenant could compromise an entire server.
A Brown University professor’s in-person makeup exam after suspected ChatGPT cheating saw scores plunge about 50%, fueling Ivy League soul-searching over how many students quietly outsource coursework to models.
@shuv1337 @benvargas @Jborgwing oh thanks!!!!
thank you for understanding it. here is my japanese article, but you know why i made the dicision...
(i'm writing a blog post to explain it, but sorry i'm so busy. i hope i'll make it a public before jarred)
https://t.co/5rsQLjfK0f
when grok cli released i checked the log(every time someone release cli i check it) but they dont store token info on logs. it is like one or two month ago?
so if you find a way to calculate it, let me know! i need ppl's help!!!
and to protect ccusage from ai slop, we auto-close all PRs by default, so if you make it let me know here!!
@shuv1337 it's decent, i don't like the integrations.json format and then it needs more around community submissions / updating data, showing tools, showing what to user to pull user data etc but it'll get better
@svankina13, @shuv1337 I think I made too many mental shortcuts here xD
I'm not talking about exiting sessions or creating readmes, I am talking about a situation where every post compact or new session gets the contents of the readme file injected once into the context.
I've been doing this with Claws, they get a lot of files but only once per context window so their personas get correctly hydrated with information.
I feel like giving the LLM the context of the repo's readme IS useful, because most of the readmes are AI slop anyway, and they don't contain user-facing information anymore, they resemble an extremely bloated AGENTS.md
New Anthropic research: A global workspace in language models.
Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.
We found a strikingly similar divide inside Claude. https://t.co/aLUPBifxth
Claude Sonnet 5 is now available in Perplexity for Pro and Max subscribers.
You can also select it as an orchestrator model in Computer. https://t.co/UktzCrUZU6
i've been using this as my default since yesterday
it's super fast and i think it's the first model of theirs that clears the bar for day to day work
also worked well with browser control and took care of some admin work for me
Grok 4.5 from @SpaceXAI is live on OpenClaw.
No OpenClaw update required, just connect your X Premium or SuperGrok subscription, select Grok 4.5 under the xAI provider, and use an Opus-class model that’s fast, low cost, and ready for agentic work. https://t.co/MMA6j2qGNy
Introducing Computer for Counsel.
Computer now connects the research databases, document tools, and matter-management systems lawyers use every day. Pull citable sources from @midpageAI, @LegalZoom, @Docusign, @netdocuments, and more.
Available for all Pro and Max subscribers. https://t.co/El3028Ua7P
The J-space lets us read, audit, and shape what Claude is actively thinking about—useful tools for keeping models trustworthy as they grow more capable. And it suggests surprising parallels between language models and our own minds.
Read the full paper: https://t.co/ge3uqY59DU
I've been letting Claude and GLT-5.5 write almost all of my commit messages recently, but I don't feel great about it
"omitting the higher-level framing needed to understand broadly what the code is doing" is definitely the key problem there
Sometimes they DO attempt to do that but guess the rationale incorrectly, which is worse than not adding no a rationale at all
I do try to link commits back to issues though, which are written by me and provide a much better answer to the reason behind a change
There is controversy in the UK for building unnecessary tunnels for their high speed trains...
...which are still vastly cheaper than California's boondoggle.