ulyssessword
No bio...
User ID: 308
ulyssessword
4hr ago
Is there some point at which you’d say “yeah, neither he nor she could have reasonably expected a pregnancy?”
As tempting as the strict liability standard is, I'd rather avoid the need to answer that question than set a threshold.
Imagine you're golfing, and your ball goes flying through someone's window. Are you responsible for that in the case where:
- You suck at golf and failed a risky shot. (Yes, you're clearly responsible)
- You're good and responsible, but flinched from a wasp sting (Yes, you're still responsible despite your precautions)
- You made a good shot, but a sudden tornado threw it away (No, that's an act of God)
- You didn't hit the golf ball, someone else did (No, that other person is responsible)
By that standard, any pregnancy short of the Virgin Mary's would be covered.
The more interesting question IMO is: Given that it happened due to your actions, what is the range of your acceptable responses? I think it does include abortion even in cases of recklessness, but you wouldn't be able to think about that in the sudden-surprise-baby framing.
ulyssessword
5hr ago
The fact that it is a possible result doesn't mean they signed up for it; we don't use that standard anywhere else.
https://en.wikipedia.org/wiki/Assumption_of_risk
From their example, if you go skiing and break a bone, the resort will argue that that's what you signed up for when you got on the hill. It's pretty normal.
ulyssessword
6hr ago
Congratulations, by driving in a car you have now signed up for a lifetime as a quadriplegic due to an accident. It was a risk you knew was possible.
Yeah, so? I also do that as a passenger of a car with a responsible driver. Or by walking down a staircase. I voluntarily take on risks every day, and live with the consequences.
Please don't do anything but accept the consequences of your choice.
Why not? That sounds reckless and/or defeatist.
I'll mitigate the risks as much as practical. If that fails, I'll minimize the consequences then deal with them as best as I can. As an example, physiotherapy is a good way to deal with the consequences of some of my actions. Is an abortion a good way to deal with the consequences of my actions? Who knows, but at least we're asking the right question now.
ulyssessword
6hr ago
Schelling points don't track base reality, they track arguments. "28 weeks" is a lot easier to argue about (why not 28 weeks + 1 day? What if it's faster-developing than average? What if they have poor recordkeeping?) than "birth".
There's also little to no developmental difference between a 17.99 year old and an 18.01 year old person, but there's a vast difference in legal status. You can argue about development all day long, but birth either happened or it didn't.
ulyssessword
16hr ago
"...Should you be in a car crash in which the victim’s life is at stake, the law does not compel you to donate blood or organs to save the victim. While it would be admirable for you to donate, you are not required to do so."
Are they sure about that? To me, it looks like the difference between manslaughter (if they die) and not (if they live). You can say that the law doesn't "require" you to save the victim, but the prison sentence on the other side of that choice doesn't make it very convincing.
The assumption that babies simply appear from the ether has weirded me out since I noticed the framing. You weren't kidnapped by a music lover and forced to give life support to another human (unless you were, in which case I support the right to abortion in cases of rape). "Where babies come from" has well-understood causality. If you sign up for baby-making, then you can't act shocked when you make a baby.
ulyssessword
1d ago
In fact, we know they're crappy because in court filings Anthropic stated they made 5 billion in total GAAP revenue between 2023 and the end of last year.
How does that follow? Taking their stated numbers literally (10x/year, $100M annualized in Jan 2024), you get:
Jan 2023: $10M annualized / 12 months per year = $0.833M in the month
Feb 2023: $0.833M * 1.21 monthly growth multiplier = $1.01M
...
Jan 2024: $8.33M ($100M annualized)
...
Jan 2025: $83.3M ($1B annualized)
...
Dec 2025: $688M
for a total of $3.94B. They put a "+" in their chart, which easily explains the 20% disparity. Their "0" in 2023 is a literal rounding error at $35M for the entire year.
ulyssessword
1d ago
IIRC, Russia hasn't declared war, so they can't draft anyone (at least without changing the law). It's all "volunteers".
ulyssessword
2d ago
Can I just point out that 21 hours seems too short for negotiations?
Maybe, maybe not. There are two types of negotiations: ones where the two sides ranges of acceptable deals overlap, and ones where they don't.
When they overlap, the name of the game is to get as much of your value without pushing down their value below their acceptable limit. This involves all sorts of tactics like setting fake "hard limits" (so that they're afraid of going below your acceptable limit, and offer better terms), hiding your true values, convincing them to change their assessments, etc. Spending extra time can get you more information about the other side, and let you get better deals.
When they don't, it's just a matter of figuring that out. This is usually complicated by the fakeouts used above, but it wouldn't need to take dozens of hours to discover. Spending extra time at that point is useless for negotiation purposes, but it might still send a political message, or serve as information gathering, or something else.
ulyssessword
5d ago
Your "based on sex" is too broad: Prejudice against men is tolerated, but prejudice against women isn't. Same with race and religion: Prejudice against white people and Christians is tolerated, prejudice against black people and Muslims isn't.
ulyssessword
8d ago
Yeah, I'm guessing he didn't do a full cost-benefit analysis before deciding on a simple strategy he could use in everyday life, but there's still a reason why he has that heuristic.
ulyssessword
8d ago
Locally, last I checked:
- glass "recycling" isn't. It's garbage that you can place in the blue bin. Beer bottle reuse is worthwhile.
- plastic recycling is mostly split between shipping it to Asia (and hoping it doesn't get dumped in the Pacific) and worse than new manufacture
- paper is kind of a waste, but I guess it's okay since a fraction of the infrastructure is paid for by the fully-useless ones anyways
- aluminum is absolutely, definitely worthwhile.
A heuristic of recycling being useless is accurate for 75% of those material categories. He's wrong about the last one, but I find it hard to blame someone for not knowing about how energy-intensive refining aluminum from bauxite is and how much easier it is to recover from cans and other scrap.
ulyssessword
11d ago
That's what you call self-documenting code. didTimeTravelHappen is there to tell you if time travel happened. Obviously.
ulyssessword
11d ago
lol, I wish. It was in Macro Script Command Language (and not even the latest version of it). I use python whenever possible at work, but still end up spending a lot of time on ladder logic, MSCL, or other strange programming languages.
ulyssessword
11d ago
The problem is that it works perfectly fine for anything between 0.01 and 9.99.
ulyssessword
11d ago
What's the stupidest computer bug you've ever made/seen?
I was having problems in a script because 0.12 < -3.45 when you compare the strings alphabetically.
ulyssessword
14d ago
There's no need for visual screenshots. Claude Code (or Codex, or whatever) could feed the model keystrokes, or diffs, or some other text-like data. It wouldn't be general-purpose, but it could cover both the IDE and the chat window.
ulyssessword
14d ago
Did you run it again on the newer, better code base, to see if it could catch another 75% of the duplications? And then again for a 75% chance to catch the last one?
ulyssessword
14d ago
There's a good amount of duplicate logic throughout for things like setting up and making the requests, caching, etc. I asked Claude to look over the code and extract out the duplicate logic into a single implementation Here's how it messed up just the authentication part of it:
Out of curiosity, what's the denominator? I'd be a lot more impressed if it missed 4/100 duplications than if it missed 4/5 of them.
ulyssessword
14d ago
Opus is like GPT, Sonnet is like GPT-mini, Haiku is like GPT-nano (although this size is largely defunct).
ulyssessword
15d ago
...motivated by exclusivity is a weird take...Surely many more people will be invited, not fewer.
That's a large majority of "exclusive" things. Exclusivity usually comes from creating something new that doesn't include most people. Creating exclusivity by kicking the existing people out is possible, but it's rare.
ulyssessword
18d ago
287, which is apparently terrible for the self-reports here.
ulyssessword
19d ago
Can you please explain what this even is?
Step 1: Install a browser (maybe Chrome?). Use ??? controls to make it the only accessible application.
Step 2: Create a plugin that blocks unwanted content (or everything but specifically selected content). Use natural language instructions to an AI instead of writing the code directly.
Step 3: Install that extension on the browser, and use ??? to make it impossible to disable.
Result: A laptop that can only do one thing. For example, watch a single channel's YouTube videos.
ulyssessword
21d ago
That would be effective, but imposing new taxes isn't "altruism" anymore.
ulyssessword
22d ago
Who's making that argument? I'm saying it's not zero, and you are the one that can't imagine anything less than infinity.
The Violinist argument essentially goes:
I reject #2 because you weren't kidnapped by music lovers in a typical pregnancy. You weren't an uninvolved bystander or innocent victim, and that causal link changes the situation. Is it enough to change the headline results? Again, who knows, but at least it's the right question.
More options
Context Copy link