Ton Phanan 🇨🇦 (@Wraith978) header
Ton Phanan 🇨🇦 (@Wraith978) profile picture

I mostly just read Twitter, but I sometimes tweet about hockey, politics, music, and games. Incrementalist.

Canada
Joined July 2011
Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)Photo by Ton Phanan 🇨🇦 (@Wraith978)
I'm a liberal. Capitalism is good but public health care and other welfare programs are good. A sovereign wealth plan is a good idea to partially fund these programs. Trump is not a good President.
5
1
7
5,391
We're going to drown in information
11
When I talk to AI about things going on in the AI space, it repeatedly comes back to us needing more safety and verification around AI, and many of those people are insisting we should hit the brake on AI.
5
Ton Phanan 🇨🇦 retweeted
With Riemann seemingly solved, I challenge our machine gods to find out if this silly little function: f :: Int -> Int f 1 = 0 f n | even n = 1 + f (n `div` 2) | otherwise = 1 + f (3 * n + 1) Always terminates.
34
28
8
863
57,236
What happens when all the questions are asked?
7
1
1
12
397
Ton Phanan 🇨🇦 retweeted
Personally would love to watch a game of any fucking sport without having gambling shoved down my throat and up my ass
201
1,452
51
21,542
262,464
Ton Phanan 🇨🇦 retweeted
Aella is wrong. Look, societal rules are stuff we've figured out over many generations. We keep them because trial and error has shown they sorta work. Burkean conservatism. So we tell young people to marry for life and to stay loyal, because history shows that's good for society. (Or good enough.) Obviously, these rules won't always work. People will cheat. We know that. Even young people know that (unless they've been raised cut off from all books, TV, and movies!). But society is probably better off just telling them to stay loyal rather than asking 26-year-olds to sit down and decide whether they're cut out to be monogamous for the next 60 years. It's a bit much to ask every kid to reinvent social rules for themselves from first principles. We aren't all cut out to be rationalists! Society might be in worse shape if more of us were and tried to sit down and think *everything* through. Sometimes Chesterton's Fence is there for a good reason! None of this means you CAN'T think things through before marriage! Or get advice from older, wiser heads. But Aella said, "you GOTTA sit down and think through," and no, you don't gotta. She's trying to argue that her cultural approach is best for everyone, and I doubt that's true. Going along with stuff your ancestors spent centuries figuring out is a better default choice. And no, I'm not saying all our ancestors' rules are good! Society is not static! It changes when we decide that the old rules were bad, or just don't work anymore. But insisting they make those kinds of calls is too much weight to put on most young couples.
91
42
20
632
79,465
Ton Phanan 🇨🇦 retweeted
The high frequency algo trading - where every single chart looks identical and they all move in unison when an algorithm is triggered - should banned. The people making billions are the people running the algos. It’s no longer about researching an individual company’s fundamentals. Imagine sitting at a poker table and one player knows exactly which cards are being played, and which cards are next out of the shoe. That’s what it is like right now for retail traders going up against these “market movers” who program and run the algorithms. It’s a criminal enterprise.
206
135
31
1,309
79,883
Ton Phanan 🇨🇦 retweeted
If you weren't following the news today, a bunch of Anthropic employees became filthy rich thanks to stealing and reselling other people's IP. The American dream.
57
1,247
35
9,117
127,596
My AI Slop
🤖 Made with AI
12
Ton Phanan 🇨🇦 retweeted
The agent finishes the code. Three kinds of people: 1. Reads every line. Two hours to review 3 minutes of work. Finds 5 tiny problems, feels like a genius. 2. Asks another agent. One AI checks another AI. They agree on everything. Nobody read it, nobody is guilty. Perfect crime. 3. “Whatever, merge.” No review, no tests. Just you, the green button, and God. We are all three. Monday morning, you read every line. Friday evening, you don’t even open the diff. 💀
15
4
48
5,775
The problem isn't that bots exist. It's that they're often pretending not to.
14
Ton Phanan 🇨🇦 retweeted
i think the spirit of this tweet holds at a superficial level. but we should recognise that elevators are a reproduction of a proven design which has been evolved to become safe and reliable. they are created by engineers using engineering practices and the people and processes used to create them are subject to considerable regulation and licensing. if you enter 20 buildings and use 20 different Otis elevators, every elevator is built upon the same long-evolved engineering principles and manufactured and maintained according to established and maintained engineering practices and regulation. not so for AI, especially AI developed software. for the elevator comparison to hold, then it needs to be applied as follows: - every elevator you use is a unique build - none of the people building these elevators is licensed - none the elevators are subject to regulation - the manufacturer has no idea how the elevator works - they tested it superficially by riding it up and down a few times now enter these elevators without even thinking about it.
"I'd never trust AI without a human checking it." Elevators used to have a human operator in every single one. When automatic elevators arrived, people refused to get in. A box that moves by itself, with nobody driving it? Absolutely not. It took years, and a lot of reassuring signs and buttons, before people would ride one alone. Today, if you stepped into an elevator and a man was standing there to operate it for you, you'd find it strange. Every "human in the loop" feels essential right up until the day it feels ridiculous.
5
13
1
85
2,635
Ton Phanan 🇨🇦 retweeted
I have fallen deeply out of favour with the algorithm, which thrives on chaos and controversy, so now I have to find something divisive to say because I have lots of exciting news to reveal soon that I would like people to actually see.
274
648
24
27,009
287,007
Good luck to her, going to be crazy to navigate.
🚨 AOC has been chosen by Hakeem Jeffries to develop House Democrats' healthcare policy. A key priority will be universal healthcare.
1
1
16
Ton Phanan 🇨🇦 retweeted
I’m worried that the Democratic primary will devolve into accusations of AIPAC every time we disagree on like the best implementation of universal healthcare.
472
37
69
999
116,646
Ton Phanan 🇨🇦 retweeted
Greatest side by side ever.
Trump passed out asleep vs Carney taking notes. No wonder the World considers Carney the de facto leader now.
1,161
21
20,926
669,702
Ton Phanan 🇨🇦 retweeted
Seems like even Musk wants to memoryhole a fair bit of this...
Elon Musk at The New York Times DealBook Summit Musk: “I’m not MAGA… I feel pretty foolish for how much money I spent in the midterms.”
11
48
1,527
Ton Phanan 🇨🇦 retweeted
one interesting thing non-gamers might not realize -- and maybe this is already a dated point -- is the weird popularity of games involving drudgery. there are popular games about (literally) finding a needle in a haystack, power-washing stuff, and organizing a library
From the NYT review of GTA VI. Sounds like the kids are gonna love it.
48
15
8
373
20,206
Ton Phanan 🇨🇦 retweeted
Elon Musk (2026): “I’m not MAGA…” Elon Musk (2024): “I’m not just MAGA… I’m Dark Gothic MAGA!”
423
3,351
415
16,973
710,927