
@KyleOlneyi
iAccount based inUnited States
About this account
- Account based in
- United States
- Connected via
- United States Android App
Account-level information from X, not a live location or the device used for a specific post.
Philosopher, Activist, Consultant, Engineer; Let's change the world together.
San Francisco
Joined June 2011
- Tweets10.9K
- Following4.5K
- Followers480
- Likes49.7K
Kyle Olney retweeted
Things we learned recently:
-Data centers are bot farms
-Bot farms sway elections
-Billionaires want data centers
Do the math.
💯 - the unspoken truth is that you have to lie about current capabilities to close deals & then turn around and build whatever you were able to sell.
This is the cause of tension between sales & prod/eng depts.
🎯 - this is why it feels like society is about to collapse btw.
All the boomers are trying to hold their breaths long enough to cash out + run with the bag, while all the under 40yo's are desperately trying to scrape enough together before the ceiling caves in.
Bad situation..
Kyle Olney retweeted
No, you cannot walk 10,000 steps daily, get 8 hours of sleep, cook every night, clean every day, take care of a family, make time for your own hobbies, and still be productive at work every day. This is not just propaganda, it is nonsense. Free yourself from it.
Kyle Olney retweeted
So if I purchase a DVD and digitize it, that’s copyright infringement. But if an AI company buys my book and scans it to train their system, that’s perfectly fine? Once again, corporations are allowed to play by their own rules. I’m so sick of the hypocrisy.
Wake up Neo...
We’ve built an entire civilization around the idea that humans must sell most of their waking hours just to survive.
And somehow we’ve convinced ourselves that this is natural.
It’s not.
For thousands of years, technology has been about freeing us from necessity.
AI and robotics could finally take that idea to its logical conclusion.
A world where existence doesn’t have to be earned through labor.
Umm....yes, obviously.
Multiple years of net 0/- GDP growth when you strip out AI infra spending while rates surge & the underlying economy deteriorates is an exact manifestation of Dutch disease. Made much worse by how financialized the US economy is.
We're all in on tokens
very strange system where a 25 year old who can’t afford rent pays the retirement for someone living in a $20 million home.
Kyle Olney retweeted
The whole point of this circus is to ascend. Earth gave you trauma, lust, ego, money, fear, exes, taxes and algorithms, and you still think the mission is to get comfortable? Fam, how many side quests do you need before you clock the main one?
Earth is basically spiritual GTA. Temptation everywhere, random boss fights, people trying to rob your energy, and half the population still thinks the mission is get a mortgage and die politely.
The game is simple. Master yourself, stop getting baited by every impulse, learn what the fuck you are, then ascend. People turned the tutorial into a 70-year side quest.
Imagine reincarnating 12 times just to keep falling for lust, ego, money and people’s opinions. Brother, complete the fucking level.
You got dropped into a body with desires, fear, trauma, hormones, ego and free will, then surrounded by distractions 24/7. And people still can’t work out there might be a test going on. Comedy.
Life keeps sending the same lesson wearing different faces and people keep calling it bad luck. Fam, at some point you’ve gotta pass the fucking exam.
The wildest part is people think the goal is comfort. Nothing about this place looks designed for comfort. It looks like a spiritual obstacle course
You came here naked, leave with nothing, and somehow people think the whole point was collecting shit in between. Mad
In light of the latest Anthropic's latest not-particularly-oblique attempt to build a case against open weights, I thought I would write down how I currently think about open weight AI models.
(1) Open weights have been deeply, irreplaceably useful for AI safety.
A careful observer will note that these growth numbers track suspiciously close to "real" GDP growth in their respective economies.
That is not a coincidence. Prosperity is a direct outgrowth of increasing per capita energy consumption.
We need A LOT more of it in the 🇺🇲
Few metrics are a better read on real economic growth than electricity production.
US’s was flat for 20 years😳
US is finally reacting; bond investor consensus still thinks the US can catch up without the bond market being crushed in real terms, but that’s not possible. #YCC
In case you're wondering why everyone is now *super interested* in building data centers...
These tech companies are all grifters The American taxpayer is funding the entire operation.🚨. most sophisticated financial extraction scheme in American history… look at the data they use government tax subsidies to mask operational losses, then use the appearance of profitability to justify massive valuations, and use lobbying to prevent any regulation that would expose the scheme.🚨👀
Total tax reduction across the four companies:🚨👀
• Meta: $6.8B reduction
• Microsoft: $11B+ reduction
• Amazon: $8B reduction
• Alphabet: $7B+ reduction
• Total: approximately $32.8 billion in tax reductions
The tax subsidy scheme works like this: The companies spend billions on data centers (Meta spent $72 billion). The "one big beautiful bill" lets them immediately deduct this spending from their taxable income. So they're using government tax breaks to build the infrastructure that gives them monopoly power, while simultaneously lobbying to prevent any regulation of that infrastructure.
🚨👀
circular financing scheme. Here's how it works:
Step 1: Companies contribute to Trump's inauguration and spend on lobbying ($1M each + lobbying)
Step 2: Republicans pass the "one big beautiful bill" with AI tax subsidies
Step 3: Companies use the tax subsidies to immediately deduct billions in AI/data center spending
Step 4: This dramatically reduces their federal tax payments (Meta: $6.8B reduction, Microsoft: $11B+, Amazon: $8B, Alphabet: $7B+)
Step 5: The reduced tax payments increase the federal deficit
Step 6: The increased deficit creates pressure to cut federal spending elsewhere (social programs, regulation enforcement)
Step 7: The companies use their increased profits (from tax savings) to spend more on lobbying ($900M in 2026)
Step 8: The lobbying prevents regulation and ensures the tax subsidies continue
self-reinforcing loop. The companies are literally using taxpayer money to prevent regulation of themselves.
—
Meta:
• 2024 federal income tax: $9.6 billion
• 2025 federal income tax: $2.8 billion
• Reduction: $6.8 billion (71% reduction)
• Capital expenditures in 2025: $72 billion
• If they can immediately deduct even 50% of that ($36 billion) at the 21% corporate tax rate, that's $7.56 billion in tax savings
• So the tax savings ($7.56B) exceed the tax reduction ($6.8B), meaning Meta's actual tax reduction is almost entirely explained by the AI tax subsidies
Microsoft:🚨
• Federal income tax expense dropped by over $11 billion from FY2025 to FY2026
• Microsoft's capital expenditures for AI are estimated at $50-60 billion annually
• At 21% tax rate, immediate deduction of $55B would save $11.55 billion
• This almost exactly matches the $11 billion+ tax reduction
Amazon:🚨
• Federal income tax payment dropped by nearly $8 billion from FY2024 to FY2025
• Amazon's capital expenditures for AI are estimated at $40-50 billion annually
• At 21% tax rate, immediate deduction of $40B would save $8.4 billion
• This almost exactly matches the $8 billion tax reduction
Alphabet:🚨
• Combined current federal and state income tax expense dropped by over $7 billion
• Alphabet's capital expenditures for AI are estimated at $35-40 billion annually
• At 21% tax rate, immediate deduction of $35B would save $7.35 billion
• This almost exactly matches the $7 billion+ tax reduction
—-
Total across all four companies:
• Total tax reduction: $32.8 billion
• Total capital expenditures on AI/data centers: $200 billion annually
• Total tax savings from immediate deduction at 21%: $42 billion
• The numbers match almost perfectly
This means the AI companies' tax reductions are almost entirely driven by the "one big beautiful bill" tax subsidies, not by any actual business improvement.🚨👀
Kyle Olney retweeted
ok i think i finally figured out why alignment has gone basically nowhere after 15 years and several billion dollars. the problem is hard sure. but have you met the people in charge of it
1) the leadership does not know how computers work. ask one what happens when you type a url into a browser and watch them start sweating. dns is a spooky word. they genuinely believe the model lives in "the cloud" like a ghost haunting a castle
2) zero idea how large scale distributed systems work. never been paged at 3am. never watched one bad bgp announcement take half the internet offline. their big threat model is the ai "copying itself across the internet" like its a floppy disk virus from 1998. brother it needs 50 megawatts, a liquid cooling loop and a personal relationship with jensen. it is not escaping onto your smart fridge
3) they dont understand the stack they are supposedly protecting us from. ask about inference servers, kv cache, batching, egress, rate limits. blank stare. ask how the model will seize the power grid and you get a 45 minute answer with a hand drawn diagram and three links to their own blog
4) nobody can define an agent. ask ten of them and you get eleven definitions and a 30 page google doc. to the rest of us an agent is a while loop that calls tools and eats a 429 every thirty seconds. to them its a digital god in its larval stage. my agent cant book a dentist appointment without asking me three times if im sure
5) they started from "everyone dies" and worked backwards so every result is bad news somehow. model refuses, deceptive. model complies, sycophantic. model does well on evals, well now its scheming about the evals. i cant think of a single thing a model could do that would get them to say ok maybe we're fine
🎯
No, "anthropomorphizing" is a technical problem, about how AI works, not something like woke language policing. We say it's wrong because it's "technically" wrong not "morally" wrong.
If you fail to set the brakes when parking on a hill, and your car then rolls down the hill, you don't claim that the car "decided to drive away on its own". Such anthropomorphizing distorts what really happened -- that you failed to set the brakes.
The same applies to AI. When you "anthropomorphize" the AI you fail at understanding what actually happened.





























