satisfice

Thank you for this post. I hope we can get more people to rally around these sentiments.

A counterfeit $100 bill may be indistinguishable from a real one. A stolen $100 bill may be indistinguishable from a real one. The measurable physical mechanics of rape may be indistinguishable from consensual sex. The King of England is indistinguishable from a normal human. In each of these cases, the specialness of the item is culturally constructed by humans.

And in each of these cases, it would be catastrophic to humans if we disregarded that construct.

The consequences of treating an information construct as an entity with rights and moral standing would be so terrible that I hope people who yearn for it lose their rights.

My fan theory of Terminator: It was never the machines trying to kill humans. It was a handful of billionaires who controlled the machines attempting to kill all non-billionaires. They merely claimed they had no control.

Now I’m thinking that some of those billionaires may have worshipped the machines and convinced themselves that they were helpless. Miraculously, their own families were spared.

The burden of proof is not on us to show why wishful, magical thinking is wrong.

People who claim that AI will take over can’t point to any evidence. All they have is speculation.

It’s worse— it’s seeking to replace every single aspect of what it means to be YOU in the world. Some people are literally trying to “fire themselves” and be replaced with digital twins. Perhaps those people are independently wealthy and also have no need of human connection? For the rest of us, it is a sickening prospect.

AI is automated irresponsibility, and it is nothing like any earlier transition.

When a technology trend means people literally won’t be able to tell if you are living or dead, and also stop caring about the difference— that’s unprecedented in the history of humans.

This article is 99.9% empty space.

It’s nice to see something on HN that isn’t about writing a prompt so that you can pretend to work.

Please don’t say “okay, thanks, that’s great” unless you mean it.

She wasn’t hating it in the abstract. Read the article. She’s not upset with mathematics.

Look at that? Downvoted. Probably by no one who has ever studied testing.

That's a good one.

“Worse: it's a conversation killer. There's nothing to respond to. Your wall of text suppresses dialogue. They can't reply, can't push back, can't clarify. It's a weapon disguised as helpfulness.”

I can reply. I can push back. I can clarify. I am not helpless.

absolutely right

“Eval” is not testing. This post is written as is no one ever heard of a thing called testing.

For the uninitiated: if you think testing is nothing more than simple operations and assertions, then you don’t know anything important about testing.

I am saying that AI cannot take responsibility. My line of reasoning assumes that you understand that responsibility/accountability is a crucial element in offering products and services in society.

Although software companies like to disclaim all warranties to the extent the law allows, neither the market nor the FTC will allow you to breach contracts, make outrageous false claims, or offer products that destroy life and property.

AI cannot and will not behave responsibly. I’m not opposed to AI, but I use it as a tool that I supervise. I review and test what it produces. People and companies that fail to do that will ruin their reputations, I predict.

AI cannot take responsibility. It will not indemnify you. My decade is going just fine.

The author of this post seems to have left something out. He never clearly states his thesis.

From what I can piece together he seems to think that agentic coded projects are being routinely dismissed as mere slop, and he feels that is wrong. But I’m not sure how that connects with the argument that if vibe coding were so great why haven’t we seen a duplicate of photoshop?

I can argue that vibe coding is bad even if someone produces something with it that seems good. Not saying I necessarily want to, but I could. Just because you have “90 test cases” that pass, or that you personally are happy with your own product is not proof of the success of vibe coding. (It is evidence, perhaps, but not proof… The evidence can be debated.)

The problem of consciousness is hard partly because it is objectively hard to make conscious people— all of whom are experts at the experience of living in their own bodies— agree on what consciousness is.

A model is an idea, activity, or object that represents some other idea, activity, or object. A good model is one that helps you understand or manipulate the thing that it represents.

Then it should easy to show a world where we are all not in abject poverty. We’re waiting.

Another article that passionately devalues testing. It should be called “shrug and ship.”

But at least it dares to throw out stuff that doesn’t make sense.

If we can’t decide what constitutes evidence in the first place, then it is pointless to wonder what more evidence could possibly be found convincing.

Nothing I can say as a human proves that I am conscious if there is a possibility that I am reciting a memorized text. The presentation of a text is obviously not restricted to conscious entities.

When technology was further away from any conceivable goal post, we didn’t have to settle the question of true goals and adequate protocols. Now we do.

This is a testing issue that is built on a modeling problem that must pass philosophical muster.

I’d like to think I would destroy them wherever I encountered one, like a highly specialized Zorro. But I’m probably a coward.

This is only even an issue if your repo is unusually popular. I put repos out that I am excited about but almost no one else seems to care.

I don’t want the contributions of strangers and no one seems to want to contribute anyway. I open sourced it just to make a place for someone to find it and fork it if they please.

Are there really people out there who don’t realize its okay not to do unpaid labor for randos?

Then you will love the tisking social justice warrior attack!

Rich people will start using this to decide when to flee.

Google will use the popularity of this site as a leading indicator in its own index.

I bet it starts "Roses are red, violets are blue..."

They should investigate alternatives to solid gold carriages. Gold is not only expensive, it’s very heavy, requiring much more energy to operate. It also has less tensile strength and its way too ductile compared to aluminum.

I’m not a railway engineer, but encrusting the tracks with “pillow cut” diamonds from “actual pirate hoards” seems like a security risk. And pirate hoards are only more rare over time.

Can you do that responsibly? This is the often ignored question. Maybe the answer is yes. But what bothers me is when people just ignore the question.

In logic, if you allow a contradiction into your reasoning, you can prove literally anything (by the Principle of Explosion). Similarly, if quality does not matter, you can achieve anything else (I can fly to the moon with a quality level of zero). And if responsibility doesn't matter, any lie is as good as the truth.

If, however, lies aren't as good as the truth, then responsible behavior might matter, and then we must consider which behaviors might be expedient yet irresponsible.

Warp is a more powerful shell. It has a better UI. I almost never use its AI features.