Blano @blano

This is so insanely corrupt, I can’t even believe it. More than half the donors to Trump’s $400 million White House ballroom just won over $50 billion in new federal contracts in six months. And here’s the part that should make your blood boil. Sixteen of these 27 donors were facing federal enforcement actions, antitrust reviews, labor cases, securities charges. Many of those cases have been quietly dropped or scaled back since Trump took office. You write a check, your legal problems disappear. That’s not a coincidence. The White House won’t even release the full donor list. They’re hiding it on purpose, because daylight is the one thing pay-to-play can’t survive. A federal judge already ruled ballroom construction has to stop until Congress authorizes it. Government is supposed to serve the people, not auction itself off to the highest bidder. When access goes to whoever pays the most, working families always end up paying the price. We either end the corruption, or the corruption will end us. wapo.st/3QmJjSz

These are extraordinary photographs. I also love the power of black and white in prints.

Joseph Miscimarra 🇺🇸 @jmisci

4 days ago

This week I came across the obituary of a photographer named David Plowden. I was unfamiliar with his work, but decided to browse his website after reading that he specialized in photos of trains and industry. I’m not much of an art guy, but these photos are astonishing. (1/4)

keyboard_arrow_left Previous keyboard_arrow_right Next

86 884 5K 104K 1K

@hthieblot I agree with others on WinZip. Also, I definitely used AOL IM even after escaping the AOL CD accounts (yup, I was on in 1996). Also, I was a heavy Eudora user for email.

LOVE the creativity!

Om Patel @om_patel5

5 days ago

THIS GUY LIVES UNDER SFO'S TAKEOFF PATH SO HE BUILT A CEILING PROJECTOR THAT TRACKS EVERY PLANE FLYING OVER HIS HOUSE IN REAL TIME he uses a cheap $30 radio receiver to pick up the signals that planes broadcast while flying. then projects them onto his ceiling in real time

374 2K 27K 3.0M 8K

Are AI agents "colleagues" or just "workflows"? Top COOs at the Fortune Summit couldn't agree. But for risk and compliance, both sides miss the mark. The solution: treat AI agents as digital employees under a hybrid governance framework. 🧵👇 Treating AI as a peer "colleague" backfires. HBR & BCG research shows it causes human workers to shift accountability, review work carelessly, and scapegoat the AI. But treating AI as just a "workflow" ignores its dynamic, decision-making agency. The answer: treat them as digital employees. They aren't peers, but they need an HR spectrum overlay alongside IT asset management. IT manages the compute, licenses, and token budgets. HR processes manage the behavior and roles. What does an HR overlay look like for AI? • Written job descriptions & boundaries • Onboarding protocols & policy alignment • Regular performance reviews (auditing for drift) • Formal deactivation (offboarding) Crucially, every digital employee must report to a human manager. If the agent fails, the manager is responsible. By applying an HR lens to AI, compliance leaders can keep lines of human accountability crystal clear. 🧵 The Fortune piece: fortune.com/2026/06/02/sho…

This will save a ton of time!

Josh Woodward @joshwoodward

2 weeks ago

A top feature request is rolling out in @NotebookLM: Google Drive files will now automatically sync! 🔄 We're actively rolling this out, starting with 10% today and ramping up soon.

100 135 1K 124K 200

@mpopv What do you see when you stare into this? ;-)

Governance is not an innovation brake. It is the brake system that lets a Formula 1 car run at 200 mph. Without a Digital Employee framework, you are stuck in low-stakes Shadow AI pilots that never touch core revenue workflows. Three actions for boards this quarter: * Digital Workforce Registry with named human principals * Role-specific Agent IDs (kill the shared API keys) * "Audit of Intent" logs for high-stakes decisions and make it an active framework, not just a slide. Full brief: blanewarrene.substack.com/p/ai-agents-ar… #AIGovernance

Brilliant case study showing the power of context at a glance. The Evolution of a Pitch Deck open.substack.com/pub/speedrun/p…

It is a stunning number, even though it is early in the adoption curve. I see this at the micro level, including discussions of simple best practices such as managing context and MD files to ensure efficiency. The meta-lesson: treat your AI assistant like a new senior hire. Ask it how it works, what it's good at, and where it struggles. You'll learn more in five minutes of conversation than in five hours of documentation. Then operationalize that internally.

Brandon Carl @brandonjcarl

2 weeks ago

Very interesting data across 2,444 companies. 82% of tokens are spent on AI-generated bugs, rework and review friction.

94 456 3K 636K 1K

I logged a meal by talking to my Mac Mini last night. It was in my food log before I sat down to eat Same AI I use in product work, different half of my life. 5 hours of build time, $0.10/month to run vs $34 for three apps before this. Speak a meal → Claude figures out what I ate → USDA fills in the calories and macros → row in a Google sheet Post is up: blanewarrene.substack.com/p/why-i-built-…

Your CLAUDE.md file is probably growing every week. Pawel Huryn's has six items and hasn't changed in months. His file contains a 2-3 sentence project description, a file structure map so the agent never scans blindly, identity context (role, audience, goals), knowledge routing that tells the agent where domain-specific files live, workflow pointers, and a three-line self-improving prompt. That's the entire file. Six items. A routing table. Writing style rules live in a voice file. Good and bad examples live in a patterns file. Platform instructions live in platform-specific folders (x/, linkedin/). Historical metrics live in a metrics file. Hypotheses and confirmed rules live in domain-specific knowledge files. Detailed procedures live in skills folders. The agent reads the routing table, identifies the domain, and loads only what it needs. The three-line self-improving prompt is the part you can paste into your own CLAUDE.md today: 1. Before starting any task, review existing rules and hypotheses for this domain 2. Apply confirmed rules by default to your work 3. After completing work and receiving feedback, update rules and hypotheses based on what you learned Three sentences. Works for any domain: marketing, testing, strategy, hiring, content. The agent reviews what it knows, applies what's confirmed, and updates its knowledge after every session without you rewriting instructions. Over time, the system auto-discovers hook patterns that correlate with engagement, voice archetypes per platform, sentence structures that get quoted and shared, and which format/length/topic combinations drive reactions. All organized by domain, all ranked by evidence. The compound gap between someone running this and someone prompting from scratch doubles every month.

Aakash Gupta @aakashgupta

4 weeks ago

This guy literally broke down everything you need to master Claude: 6:07 - Why to Stop Using Chat 9:56 - Cowork vs Code vs Dispatch 18:44 - Skills and MCP Connectors 25:03 - The Skills Marketplace 29:06 - Strategy Canvas Demo 35:14 - Skill Iteration Cycle 40:46 - Why You Need

17 254 1K 258K 3K

That is stunning @AJamesMcCarthy! I have a personal favorite I took on my iPhone on a whim while landing in San Diego. What are the odds that the shadow fits between the moving cars? Pure luck.

Andrew McCarthy @AJamesMcCarthy

4 weeks ago

I took 1.7 million photos over 6 days to catch this photo of a commercial jet in front of the sun. The moment it happened, TWO floating prominences were visible, making this not just my best aircraft transit photo, but one of the luckiest of my career! Videos of the transit 👇

1K 4K 34K 1.6M 3K

Lord, do I need this!!!!

Elí Guerrón @eliguerron

a month ago

When I was at Apple, I loved working on micro interactions that you see all over the OS. Now that I’m not an apple I still like to solve for these little problems that really annoyed me. In this case, I designed a backspace button with a speed controller, so by just pressing it

232 210 6K 529K 1K

@davidaxelrod Love the pop-up jams, be they organic or staged!

@Ike_Saul 🤦‍♂️

This is fantastic work

@

57 years ago

0 0 0 0 0

I concur - loved the crisps in the UK (wayyy better than US potato chips - though I love a wide array of US food!). Now, if Anthropic is buying AI Chips, let's get them the right recipe for AI Cod, then we can get a proper meal out of Claude Code.

“paula” @paularambles

a month ago

they call them crisps there

176 1K 22K 590K 680

Apple ends up looking patient and wise, waiting for the weight of this untenable spending to collapse on itself.

Gary Marcus @GaryMarcus

a month ago

Sheer insanity. Amazon, Google, Microsoft, and Meta collectively are spending more money than the Manhattan Project *every single month*. More than 12x the Manhattan Project every year. And what they have got to show for it? None are making major profits on AI; none has a

364 1K 5K 282K 840

no professional economist has ever been or will ever be again this funny on purpose

First To Hear It @firsttohearit

a month ago

REPORTER: You mentioned that, staying on as Fed governor, you intend to keep a low profile. Could you give us a little more detail on what that looks like? JEROME POWELL: *ducks down*

207 2K 32K 4.1M 3K