Joined July 2017
2,094 Photos and videos
the trades labor shortage is real, and it is possible to make a comfortable living in the construction industry - the six figure incomes are there. the problem is how long it takes to access them for americans. first you have to put up tens of thousands of dollars to go to trade school. then you have to spend years on slave wages as an apprentice and journeyman, before finally getting access to those opportunities. we don't have years to address the shortage, and americans shouldn't have to wait so long to make good money building up the country's infrastructure. we need trades reform now!
Narrative violation: according to Blackrock, the U.S. does not have enough workers to meet surging demand for the construction boom. AI is creating too many jobs! axios.com/2026/01/16/blackroโ€ฆ
5
1
16
6,603
young men in america rightly feel like being told to go learn a trade is yet another scam, because when you look at the economics, it is. meanwhile, there are tons of outfits who would be happy to pay six figures or more if they could actually get the labor they needed.
1
1
8
2,745
it's the same thing as the electric grid all over again. the capital is there. the will is there. there's processes and regulations in the way that aren't effectively responding to demand. let us pay to get more young men into the trades! we want to pay them six figures and get an electrician in twelve months instead of six years, but there's no way to do that today!
1
1
8
1,947
datacenters for hyperscale ai are a fundamentally new manufactured object, representing a never-before reached apex of industrial capacity and civilizational coordination. turning them into neoclassical monuments would be a mistake -- a return to neoclassical architecture would be something like a fourth-hand rehash of greek forms (via rome, then 18th c. neoclassical architecture). if they are to be symbols of progress and an american future, a new aesthetic synthesis is required. we should not be recapitulating previous forms - including the uniquely american monolithic art deco. yes we need an architectural style for our buildings that is both popular and beautiful, but we cannot allow the history of american industrial architecture to end with LARP.
We should insist that all data centers that are built are architecturally beautiful in the neoclassical style
7
3
53
8,387

"Life in Revolution is camp life. Personal life, institutions, methods, ideas, sentiments, everything is unusual, temporary, transitional, recognizing its temporariness and expressing this everywhere, even in names. Hence the difficulty of an artistic approach. The transitory and the episodic have in them an element of the accidental and the accidental bears the stamp of insignificance. The Revolution, taken episodically, appears quite insignificant. Where Is the Revolution, then? Here lies the difficulty. Only he will overcome it who fully understands and feels the inner meaning of this episodic character and who will reveal the historic axis of crystallization that lies behind it. โ€œWhy do we need solid houses?โ€ the sect of Old Believers used to say. โ€œWe are awaiting the coming of Christ.โ€ Nor is this Revolution building solid houses, but instead, it makes removals, concentrations and barracks. The character of the temporary and of the barrack lies on all its institutions." -Trotsky
8
2,081
anton ๐Ÿ‡บ๐Ÿ‡ธ retweeted
increasingly bullish on anthropic
13
16
447
92,615
i suspect 2025 was the last โ€˜normalโ€™ year; the magnitude of the compute buildout in 2026 & large scale hbm deployments (b200/tpuv7) unlocks the possibility of efficient reasoning post-training on frontier-sized pretrains. simultaneously, verifiable environments across many domains will proliferate (i expect a gold rush here as with expert labeling in the past). increasingly, models will โ€˜just workโ€™ on a variety of tasks. the next step will be for the models to learn to integrate themselves. we may not get this in โ€˜26 but we should expect progress to continue to accelerate.
RL Environments and RL for Science Data Foundries Multi-Agent Architectures Worker Automation RL as a Service Anthropic's next big bet GDPval Utility Evals Computer Use Agents LLMs in Biology Lab Procurement Patterns Platform Politics Access newsletter.semianalysis.com/โ€ฆ
13
20
276
34,287
anton ๐Ÿ‡บ๐Ÿ‡ธ retweeted
RL Environments and RL for Science Data Foundries Multi-Agent Architectures Worker Automation RL as a Service Anthropic's next big bet GDPval Utility Evals Computer Use Agents LLMs in Biology Lab Procurement Patterns Platform Politics Access newsletter.semianalysis.com/โ€ฆ
13
30
233
151,568
happy new year!
2
14
1,634
anton ๐Ÿ‡บ๐Ÿ‡ธ retweeted
30 Dec 2025
I am reminded of this 1897 account of a visit to Nobelโ€™s Scottish dynamite factory, where everything depended on a man sitting on a one-legged stool (so he couldnโ€™t fall asleep) whose job was to watch a thermometer to ensure the chemical never got too hot: โ€œThe surroundings are rather trying to sensitive nerves - your life depends, at every moment, upon a thermometer and a man on a one legged stool. Great is the thermometer at Ardeer! Nitroglycerin, a teaspoonful of which would blow you to fragments, surrounds you in hundreds and thousands of gallons. It is making itself in huge tanks, gurgling merrily along open leaden gutters, falling ten feet in brown waterfalls, so to speak, into tanks of soda solution, and bubbling so furiously in other cylinders, through the in-rush of cold air from below, that it seems to be boiling. It is being draw off from large porcelain taps like ale, poured into boxes, and rattled along tramways - all these processes proceeding as rapidly as if it were ordinary olive-oil instead of the deadliest explosive known to man. The product would suffice, if it were to explode, to cut you off in the beauty of your youth. Death instantaneous and pulverizing, encircles you, in fact, by the ton; but the man and the thermometer surround you also. The manโ€™s eyes never leaves the instrument.โ€
27 Dec 2025
We are hiring a Head of Preparedness. This is a critical role at an important time; models are improving quickly and are now capable of many great things, but they are also starting to present some real challenges. The potential impact of models on mental health was something we saw a preview of in 2025; we are just now seeing models get so good at computer security they are beginning to find critical vulnerabilities. We have a strong foundation of measuring growing capabilities, but we are entering a world where we need more nuanced understanding and measurement of how those capabilities could be abused, and how we can limit those downsides both in our products and in the world, in a way that lets us all enjoy the tremendous benefits. These questions are hard and there is little precedent; a lot of ideas that sound good have some real edge cases. If you want to help the world figure out how to enable cybersecurity defenders with cutting edge capabilities while ensuring attackers can't use them for harm, ideally by making all systems more secure, and similarly for how we release biological capabilities and even gain confidence in the safety of running systems that can self-improve, please consider applying. This will be a stressful job and you'll jump into the deep end pretty much immediately. openai.com/careers/head-of-pโ€ฆ
36
125
1,387
137,613
every few days i cycle between exhiliration and vertigo x.com/atroyn/status/17582960โ€ฆ

if anything, i lean into it much more. there is a very distinct feeling that i can isolate now when i think about things holistically, that something really unprecedented is happening. this time it really does feel different.
2
1
29
12,845
i have the distinct impression that almost everyone on earth is sleepwalking and doesnโ€™t realize it.
3
20
2,211

If youโ€™re paying attention, itโ€™s overwhelming
4
19
2,217
itโ€™s very simple. if a us politician or public figure makes a statement which, if true, would benefit china, they are a chinese asset. if they do this for free theyโ€™re a particularly stupid chinese asset.
1
19
1,738
anton ๐Ÿ‡บ๐Ÿ‡ธ retweeted
Thanks to everyone who came to The Gentle Romance book launch last week! Itโ€™s great to have it out. Also, signing books is surprisingly fun. If you havenโ€™t already, thereโ€™s still time to order copies for xmasโ€”link below.
5
3
137
62,247
nolan's 'the odyssey' has a lot to criticize in the trailers but 'historically accurate costumes are the only way to portray the western canon' isn't what people did historically. bible scenes in medieval / renaissance paintings show people in medieval european settings.
1
2
13
2,582
don't get me wrong here, turning everything into capeshit is awful and i don't look forward to nolan's trademark midwit's idea of 'smart movie' but the costumes really aren't an issue here
1
7
1,136
i think the costume discourse is mostly an artifact of collapsing consensus where people are reaching for any kind of certitude, so historical accuracy might be able to provide that, but
4
1,027
peace on earth, goodwill toward men
1
1
15
1,318
america is going to have to relearn industrial capitalism.
3
15
1,333
j.p morgan mindset
2
13
1,223
acab (assigned child at birth)
3
11
1,310