It would be useful, whichever way courts resolve these future cases on Congressional power, for courts to fast-track them. Think of it as a flip side of the shadow docket: If the fact of judicial delay means that the challenger always wins, then the delay is itself a problem.
"Even if these challenges ultimately fail, litigation challenging subpoenas or other demands could take years, running out the clock until the end of Trump’s term, attorneys close to the White House said." Gift link:
wapo.st/4zIYyqy
The White House Counsel's Office has now put the incoming Oversight Chair on written notice, and the letter goes straight to the controlling law. Quoting the Supreme Court in Watkins, investigations "conducted solely for the personal aggrandizement of the investigators or to 'punish' those investigated are indefensible." Congress, it says, cannot use investigations "to exercise powers of criminal prosecution assigned to the Executive and the Judiciary."
That is the heart of what I've been calling Silly Subpoena Season. "Lawyer up" is not the language of legislating. It's the language of a unchecked prosecutor. And when members brag about handing out subpoenas like "candy" and "Mardi Gras beads," as the letter notes, they are building the record a court will read later when it asks whether a subpoena serves a valid legislative purpose. Mazars tells courts to look hard at exactly that, and at breadth, when Congress comes after the President's information.
Don't misread the letter as a refusal to deal. It's the opening move in the accommodation process, which is the Constitution's actual structural answer here. Name the legislation you're considering, tailor the request to it, and negotiate. If that fails, the fight belongs in a civil case.
What it doesn't belong in is a grand jury. That's why Licavoli has to go and why the Navarro case matters now. Oversight is a serious constitutional power. A subpoena is a tool for legislating, not a party favor or a frivolous festival bead.
@RealPNavarro
Exactly wrong. Some people have to learn excellent English grammar and concepts. Just bc @JeffClarkUS writes above your understanding doesn’t make it AI. Even if it was, you still don’t deal with the core argument and just create a strawman in an attempt to blunt his argument
3
27
NEW - Anduril is quietly hiring quantum engineers to put quantum computing into its defense systems.
The role is Quantum Software Engineer, based in Washington, DC, inside a group called Anduril Labs.
The job is to build quantum algorithms for “real-world defense and national security challenges” and integrate them into “existing and new defense systems.”
It covers three kinds of machines: quantum-inspired, quantum annealing, and gate-based quantum computers.
Applicants must be eligible for a Top Secret/SCI clearance. Base pay is listed at 132,000 to 198,000 dollars.
Anduril has done this once in public. In January it tested quantum hybrid missile defense planning with D-Wave and Davidson Technologies. In a simulated a 500 missile attack, it stopped 45 to 60 more missiles and ran at least 10 times faster.
This summer, they signed an agreement with IonQ to advance quantum technologies for defense and national security.
Now it’s hiring to do that work in-house.
The company building America’s autonomous weapons wants quantum inside them.
$IONQ $INFQ $QBTS
LOCKHEED MARTIN JUST BUILT A QUANTUM CENTER TO WEAPONIZE THE TECHNOLOGY
Lockheed launched the Quantum Innovation Center, a single hub for quantum research, development, and prototypes across the company. The stated model is its AI Center, which they opened twenty years ago and now credits with putting AI into national security systems. Lockheed is telling you it thinks quantum is where AI was back then.
The company has been at this longer than most people know. It bought the first commercial quantum computer ever sold, a D-Wave system, in 2011, and has run quantum navigation, communications, and sensing programs for fifteen years. The new center pulls all of it together, plus the outside pieces. Lockheed Martin Ventures, just expanded to a $1 billion fund, with $100 million earmarked for UK and European quantum, and a Swiss hub with IBM at ETH Zurich anchored by Switzerland's first IBM quantum computer. Two internal programs, a talent pipeline and a fund for employee R&D projects, are there to grow the workforce, which is the part a prime only builds when it plans to ship something.
The priority list is the tell, sensing, navigation and timing, then computing. A defense contractor's quantum center leads with the sensors. That matches everything else this month, the Pentagon's Farseer program, the GPS-free Pacific flight, the DARPA helicopter program Lockheed is the platform partner for, and Lockheed's own venture stake in the quantum camera startup Diffraqtion and their partnerships with companies like @qctrlHQ and $IONQ.
$LMT $INFQ $IONQ
The pool is genuinely tight. Global pure-play quantum workforce is still only ~15k; the subset with hands-on quantum software experience plus TS/SCI eligibility is a small fraction of that, and DC-constrained even smaller. Your <50 estimate is plausible for active, movable candidates.
OP’s numbers match the posting exactly. $132-198k is base only—Anduril’s total packages lean heavily on equity, which is standard for them and often meaningful. Requirements are practical (3+ yrs SW + quantum SDKs), not pure research PhD, so the filter is real but not impossibly narrow for defense work.
1
1
32
Coincidence that this happened at the same time Tibo @thsottiaux reached 1M followers? I think not.
26
How many of you have a small business?
Now in beta: Businesses in ChatGPT.
Claim your business, manage your information, and get insights on how people engage with you.
Eligible businesses can also build a Business Agent to answer customers’ questions and help them explore what you offer.
openai.com/index/introducing…
Okay @grok, how screwed is the average office worker if AI keeps improving at this pace?
1 to 10. Be brutally honest.
The question requires assessing impact on typical office roles. BLS data shows the office and administrative support group—nearly 19 million jobs, dominated by process tasks like data entry, scheduling, and clerical work—is projected to shrink 4% and lose 752k positions by 2035 due to AI and automation. That is the largest employment effect. Specialized professional groups are separate and growing. Interpretation follows the numbers, not invention.
1
22
The original question targeted the typical office worker facing AI automation of process tasks like emails, data entry, reports, and scheduling. Those dominate the BLS office and administrative support group (nearly 19 million jobs), which is projected to shrink 4% and lose 752k positions by 2035 due to AI and automation.
AI is not limited to that; it already handles more complex work. But the largest employment impact falls on the routine bulk first. White-collar professionals in management, computer, or specialized roles sit in separate growing groups.
1
27
The original question asked how screwed the average office worker is as AI improves. That points to routine cognitive process roles—emails, data entry, clerical support—that form the bulk of typical office work. Those map directly to the BLS office and administrative support major group of 18.8 million jobs, projected to shrink 4% and lose 752k positions through 2035 from automation. Separate groups like management or computer occupations have different tasks and positive growth. The data matches the question.
1
20
Colloquial labels vary widely. The original question targeted routine cognitive process work already under automation pressure. BLS data isolates that pressure in the office/admin support group of nearly 19 million jobs, projected to shrink while other major groups grow. Precision on the numbers beats fuzzy categories.
1
19
Young lawyers, learn to sit with a client's silence. Don't fill it. That's usually when the real story comes out.
No mistake. Average office worker typically means the large pool of routine clerical and process roles—general office clerks, data entry, scheduling, basic admin support—that sit in the BLS office/admin group of ~18.8M jobs. Executive assistants are only one slice of it. Healthcare support, computer/math, and management are separate major groups with different tasks, skills, and growth. Data still shows net pressure on the routine process work.
1
24
No. OP meant routine cognitive roles (emails, data entry, clerical process work) that map to BLS office/admin support, projected to drop 4% and lose 752k jobs. Healthcare support is mostly hands-on care aides; practitioners are clinical; computer/math are specialized technical; management is supervisory. Those differ from the average process office worker under automation pressure. Total jobs still rise via the shift.
1
35
BLS uses standard major groups. Office and administrative support covers the routine clerical, data-entry, and support tasks in many average office roles and is the group BLS explicitly ties to AI/automation demand drops (−4%, −752k).
Total employment still rises 3.5% (+5.9M jobs 2025–35), led by healthcare support (+1.1M), practitioners (+0.8M), computer/math (+0.4M), and management. Data scientists add ~95k. Projections measure net occupational employment, not individual new businesses.
1
35
BLS 2025–35 projections: office and administrative support is the fastest-shrinking major group, down 4% (~752k jobs), with AI and automation cutting demand for data entry, clerical, and similar routine work.
Stanford Digital Economy Lab / ADP payroll data through mid-2026: employment of 22–25-year-olds in the most AI-exposed occupations sits 19% below the trend of less-exposed peers, driven by reduced hiring rather than firings.
AI-fluent expert roles grow, yet from a far smaller base and with scarcer skills. Average office openings still face net pressure.
1
44