And on the flip side, sometimes there's no need to learn a lesson! One of my pet peeves is when people draw huge conclusions about people/things based on way too few interactions (small sample size). Sometimes someone is just having a bad day. But if it happens again and again and again, _then_ you should draw conclusions.
Agreed. Naturally, we don't always know which event is a one off ( it used to be easier prior to proliferation of internet and then cell phones ). This likely explains some of the overcorrection I see in this area as a result. I am constantly on guard in public and if someone pulls a cell to record me, I am immediately defensive.
I guess what I am saying is that it is harder to assume it is not the type of event where we don't have to 'condition' people.
Programming on a phone is a tough sell for many since typing is slower and you have less screen real estate to view/debug the code. Using an AI agent and typing only prompts makes it more compelling. You input less, and only occasionally have to edit code instead of writing everything from 0. And even with editing, typing a prompt like "separate the X logic from class Y into a new file/class" is much faster on mobile than the equivalent actions.
> the people who don't understand why some of us don't want it everywhere don't understand that distinction, or else are financially motivated to ignore it and gaslight everyone about the categorical boundaries crossed.
This is such a common fallacy that I think it should be given a name. When you believe that the people who disagree with you must either be ignorant or malicious. Leaves no room for honest disagreement or discussion. Maybe the "dumb-or-evil" fallacy?
Perhaps; but I would argue talking to many AI evangelists is a form of selection bias. Which makes the false dichotomy conclusion reasonable given the inputs, but still inaccurate given reality.
True, it's a form of false dichotomy, but I think this specific instance is particularly interesting in that it allows the holder to dehumanise their opponent to an extent, and justify lack of discussion. It's also an incredibly common conclusion in politics after people gain a somewhat superficial understanding of both sides. I wonder if it might play a key role in social polarization.
For me the strongest arguments are the ones that can argue the opponent's side as effectively as the opponent, and then show why it's weak. And that feels entirely incompatible with a dumb-or-evil argument.
>I think this specific instance is particularly interesting in that it allows the holder to dehumanise their opponent to an extent, and justify lack of discussion.
That's a wild take and a wild leap. For my own part, I see the failure or refusal to comprehend someone else's preferences, values, or boundaries as itself a profoundly human quality, even if it's a quality I don't love, rather than one which would cause me to see someone as less human.
I will admit that, when there's enough nonsense money being thrown after a vaunted object, sensible discussion can feel pointless. Prudence goes deaf amid the din of hype.
And yes, steelmanning can be highly persuasive, but not when premises are radically different enough between two parties. It's really a more productive tool to improve your model of someone else.
There have been such a large number of OCR tools pop up over the past ~year; sorely in need for some benchmarks to compare them. Would love to see support for normal OCR tools like tesseract, EasyOCR, Microsoft Azure, etc. I'm using these for some projects, and my experiments with VLMs for OCR have resulted in too much hallucination for me to switch. Benchmarks comparing across this aisle would be incredibly useful.
> This perspective relies on seeing Chinese lives as worth less than American lives.
I'm not sure I follow this. If I was to summarise GenerocUsername's argument it would be "the Chinese government is less concerned with making their economy green, and if the US begins taking an economic/influence hit to make it's economy greener, it'll be yielding an economic advantage to China, which will canabalise more global industry in a non-green way, resulting in a net worse environmental outcome." They're claiming basically a fundamental ideological difference between the countries on climate change that, coupled with a claim of zero-sum international industry, means long term environmental outcomes are better if the US is a dominant international player today.
Sidestepping the argument itself which I believe has a number of key weaknesses (as outlined by others in the comments), can you go over how you're linking that to a devaluation of Chinese lives?
I think you have to define what you mean by "less concerned." I'll take a stab at it, which is that Chinese energy use has grown 7% while US energy use has remained roughly flat. The reason I say that this only works because you devalue Chinese lives, is because Chinese energy use remains less than half (possibly even less than 1/3rd by some measures) what it is in the US per person. If the US reduced energy usage by 10% and China's grew by another 10%, it would still be the case that Chinese people relatively speaking are living in conditions that we in the US would consider extreme hardship, as a direct result of having less energy.
Essentially you're saying that the US should bully the Chinese people into increasing hardship because it's the only way to meet our climate goals.
I had tons of these "visual styles" and I remember how MS liked to patch themeui.dll and uxtheme.dll often rendering theming broken.
There were also attempts at customizing XP booting screen to achieve the perfect "it's not Windows" effect but that could easily render installation unbootable
What an interesting use case! And interesting composition.
One thing that's interesting about the AI violin cover is that I'm not sure those runs would be physically possible at that speed on a real violin. So that composition can _only_ be played digitally, I believe.
When I used to do larger more orchestral arrangements, I was constantly getting dinged by the instrumentalists that while they were theoretically musically possible, certain runs or passages were very unnatural on the instrument that I scored them on.
For a long time I really hoped that some of the more professional notation tools such as finale would add in an ability to analyze passages and determine how realistic/natural they were for the instrument that they were set to.
Not OP, but on the off chance you haven't seen this, I found the suno explorer thing quite nice. Hitting random a few times, I'll usually stumble onto something interesting. This was the first demo I heard where some of the AI tunes gave me goosebumps close to what human music does.
reply