A rose is

@sama @gabeeegoooh Next level.

Here is a manga made by ChatGPT Images 2.0 of @gabeeegoooh and me looking for more GPUs:

@KeruboSk Someone born before 1850. That’s noteworthy.

Just out of curiosity… who out there has actually talked to someone born in the 1800s during their lifetime?

@Memetic_Theory Hegel (noun) The 19th-century philosopher who figured out how to make pussy tighter. Example: “Damn, after reading Phenomenology of Spirit my walls went full dialectics — thesis, antithesis, synthesis… now it’s gripping like the Absolute.”

Hegel? The guy who figured out how to make your pussy tighter?

NFL names work. IPL names mostly don’t. NFL gives you a mascot: animal, bird, pirate, warrior. Visual, chantable, used on its own for decades. Limited dependence on the city. IPL names feel like three committees collided: •royalty words •superhero words •corporate branding The result: commentators say Mumbai, Delhi, RCB — not the given names. My ranking, best to worst: 1.Kolkata Knight Riders — dramatic nonsense, but unforgettable 2.Chennai Super Kings — absurd, yet fully committed 3.Sunrisers Hyderabad — sounds like a beach vacation, at least it’s distinct 4.Rajasthan Royals — obvious, but it fits 5.Gujarat Titans — generic 6.Punjab Kings — minimal effort 7.Delhi Capitals — sounds like a civics exam topic 8.Lucknow Super Giants — “Giants” wasn’t enough? 9.Royal Challengers Bengaluru — liquor-brand energy in team form 10. Mumbai Indians — every team is Indian. This tells you nothing. @LalitKModi How were these names picked? ​​​​​​​​​​@mipaltan @KKRiders #IPL2026

@aravind It does happen 2/3 per year

Indian oil refineries must take extra steps on security, and sabotage from the inside. India must step up surveillance and security. I got a bad hunch that the adversary (GLISCO-DS linked) can burn a refinery to increase oil price for geopolitics & derail India's economy 2-in-1.

@himganj153 Which tournament?

Two days of working with Claude with 3200 T20 league matches, and the repeated conclusion from varoius regression analyses is that bowling strength has a *slightly* (8-10%) stronger weight than batting while predicting the success of a T20 side.

Just spoke to an HHS insider & man, Eli Lilly is going nuts with the peptide wars First of all, as we speak they are pushing the FDA to reclassify the glp-1s as biologics instead of peptides so that: 1. Compounding pharmacies can’t touch them, as they’re not allowed to dispense biologics. Only large manufacturers contracted by big pharma can produce biologics 2. They can extend the patents by 10-15 years min The FDA classifies anything under 40 amino acids a peptide, over 40 a biologic & they’re pushing extremely hard on Retatrutide Retatrutide sits at 39 amino acids, technically a peptide. But the argument they’re pushing is that certain side chains should be counted as amino acids, pushing it over the peptide threshold This strategy has precedent since pharma already used it to push fertility compounds like HCG, FSH & HMG into the biologic classification, removing them from the compounding pathway Patents expire, but the classification system doesn't

Every evening, as Mohali slows down, one 70-year-old chooses to show up with warmth. Near Sectors 54–55, on the road towards Franco Road, he quietly serves free chai to anyone passing by his home. No banner, no announcement. Just a simple act of seva that has continued for over a year. Winters bring steaming cups of tea. Summers bring both chai and cold drinks. But what truly stays constant is the feeling he offers, that someone cares enough to pause for you. In a world chasing bigger things, he reminds us that sometimes, the smallest gestures leave the deepest impact. Credits : harsimar_daily_vlogs on IG #GoodNewsIndia #Kindness #Inspiring #Punjab [Mohali, Punjab, Kindness, Inspiring India, Good News, Positive News]

@JohnConnor_US @Handre I benefited from this program. Got rid of an old voyager and got a new honda fit using the incentive.

@Handre I voted for that program and I’m disgusted to see this flagrant misrepresentation of the laudable policy goals we achieved. These cars polluted more. They were less safe. They were driven by poor people who don’t even have jobs to get to. We fixed that. Shame on you!

Cash for Clunkers destroyed 690,000 functional vehicles in 2009, creating an artificial scarcity that rippled through used car markets for over a decade. The Obama administration sold this $3 billion program as environmental salvation and economic stimulus, but any free market economist could predict the real outcome: massive wealth destruction disguised as progress. The program forced dealers to pour sodium silicate into engines, permanently destroying cars that poor families could have afforded. Politicians eliminated the bottom tier of the used car market overnight. Suddenly, a reliable $3,000 Honda Civic became a $7,000 Honda Civic (if you could find one). The supposed beneficiaries — working-class Americans who needed affordable transportation — got priced out entirely. Government intervention always creates unseen victims, and Cash for Clunkers delivered them by the millions. Single mothers, college students, and minimum-wage workers watched their mobility options vanish as used car prices soared 30% between 2009 and 2014. The environmental gains proved negligible too: most clunkers averaged 15-17 MPG while replacements hit 24-25 MPG. Destroying half a million cars to improve average fuel economy by 8 MPG represents the kind of central planning that would give Soviet bureaucrats a hard-on. The wealth destruction extended beyond sticker prices. Higher transportation costs forced people into longer payment terms, creating a debt cycle that persists today. Cash for Clunkers normalized 84-month auto loans, turning cars from depreciating assets into multi-year financial anchors. Bureaucrats congratulated themselves for moving inventory off dealer lots while condemning an entire generation to transportation poverty.

@pmarca guilt-free instinctual living? Easier said than done when others depend on you to pay the bills. Could work in small doses.

Many wise people are saying.

@growing_daniel I see many more posts about TBPN and none from the show itself.

All this talk about TBPN exposure and their daily show stream had like 150 viewers whenever I dropped in on it

@mattiapomelli Opus 4.7

Opus 4.7 vs Gemini 3.1 Pro vs GPT 5.4 Same prompt. Which one do you prefer?

I built my first computer in 1967 out of discreet transistors. I got my first email address in 1979. I’ve seen some things…

List of recent related papers by economists. Definite productivity gains. Unclear impact on jobs. Brynjolfsson, Li, Raymond GenAI at Work. Customer support, real workers. +15% productivity avg. Biggest gains for the least experienced. The anchor empirical paper. Bick, Blandin, Deming Rapid Adoption of GenAI. 23% of workers used AI for work in the prior week. ~1–5% of work hours already AI-assisted. Fastest tech diffusion on record. Humlum & Vestergaard LLMs, Small Labor Market Effects. Danish admin data. Real adoption, real task restructuring — but near-zero effects on earnings/hours after 2 years. Best corrective to hype. Cruces et al. Does GenAI Narrow Education-Based Productivity Gaps? Without AI: 0.55 SD gap. With AI: 0.14 SD. AI closed ~75% of the education-productivity gap. Striking. Yotzov et al. Firm Data on AI. ~6,000 executives, 4 countries. 70% use AI. 80%+ report no productivity impact yet — while expecting gains ahead. The adoption-impact lag in one stat. Acemoglu Simple Macroeconomics of AI. Aggregate gains depend on how much task space AI actually affects + size of cost savings. Useful framework against vague trillion-dollar claims. Autor Applying AI to Rebuild Middle Class Jobs. The normative counterweight. AI should augment expertise and broaden access, not just automate and concentrate rents. Hui, Reshef, Zhou GenAI & Online Labor Markets. Workers in AI-exposed occupations saw real drops in employment and earnings post-ChatGPT. Best evidence of displacement pressure in gig markets.

He is so wrong on this. AI will certainly cause dislocation, but like all technology it will also create new jobs and opportunities in the medium term. AI and robots will also not produce goods and services in excess of money or demand that there will be no inflation. Both of these are classic mistakes made by those who think that there is a finite number of jobs to be done in the world and a finite set of consumer demands. By their logic, we have already exceeded everything that even the wealthiest person could have imagined in 1800, so there should be no jobs or inflation in the 21st century. By the same token, @elonmusk 's universal high income will bankrupt any government that attempts it.

@simonw I am having little to no luck in getting LLMs draw a basic geometry diagram of which there must be literally thousands in the training set. Deeply puzzled. Prompt below. explain visually using a diagram. A central angle stands at the center. An inscribed angle stands on the circle. A key theorem is this: An angle at the center is twice the angle at the circumference standing on the same arc.

@Zigmanfreud @grok Sports is very hard for generic AIs. Maybe out of reach. Specialized engines front ended by AI will likely be the solution. Grok and other general AIs are completely clueless about cricket, for example.

Why I’m not very impressed with the ability of AI to “think”… On the 40th anniversary of Jack Nicklaus winning the 1986 Masters, I asked @Grok Imagine to take the iconic image of Jack making the putt on the 17th hole, and turn it into a video. It was very clearly clueless… 🤷‍♂️

Excavated a 33 year old code base, applied minimal compatibility mods and pushed it to github. I wasn't able to (or didn't bother to) easily share this in 1993. I hadn't seen this for decades and LLMs (ChatGPT pro in this case) made this process painless. github.com/nsram/CLOPT

@davepl1968 The algorithm hasn’t yet figured out that your content is unique, original and authentic.

My demographic is pretty centered amongst 50-year-old male tech enthusiasts. In fact, my YouTube stats are literally 99% male (apparently, I'd be an otter according to the comments). Yet I have 1M subs on YT, 100K on X, and only 12K on Facebook! Completely backwards from what you might expect. Is it significantly harder to build a presence on X? I rarely check, but it looks like I lose more followers due to my random hot takes than I build from content, so it's kind of a wash! Either way, you know I'm cashing that sweet, sweet $482 check every month! That's like a nice dinner out with the wife!

@elonmusk This is excellent content. From a olde wizard.

Yes