I recommend reading Yuval Noah Harari's Nexus for a deep discussion around this.

He makes the point that what makes this AI age much more dangerous for mass surveillance isn't just the collection of data, which has indeed been possible for a while, but the new ability to have AI sift through that enormous volume of information, an ability which until recently has not been possible in a meaningful way without a ton of manual work to support it.

Older attempts at mass control of a population already involved mass surveillance, even in a large amount of detail, but even when capturing in detail all citizens' activities, there were just not enough people around to be able to dig through that and analyze it. This has been somewhat true even with the help of computers, though computers have certainly already been making this easier.

But now you can just give all that data to an AI with your instructions, and it'll apply some sort of "judgement" on your behalf, completely autonomously, and even perform actions against those folks it finds, again autonomously, without needing to manually build a whole infrastructure to do that with manual rules. That's a very meaningful upgrade for someone wanting to control a population.

Comments URL: https://news.ycombinator.com/item?id=47186416

Points: 1

# Comments: 0

Something can said to change from a certain standard even if it wasn't perfectly constant to begin with. For example, if I always kept my house at 65-75 degrees for the past year, and now it's 85 degrees inside, I could certainly say that the temperature in my house recently changed and gotten warmer. That might lead me to check whether my AC's working, rather than say "well I guess the temperature has never really been constant, and 85 is within the range of possible non-constant temperatures, so everything's perfectly normal and nothing has changed."

It makes it really easy to work with stacked PRs and keep them in sync, and do some nice tricks like decide which part of the stack a commit goes into, and apply and unapply parts of the stack.

I loved the operation log and how you can instantly revert to an older operation, as an escape hatch from weird merges and other things that go wrong.

But. It is absolutely incapable of dealing with merge conflicts. No matter how trivial the conflict, GitButler makes you click through sometimes multiple iterations of a rebase-like flow where you fix an issue and then go back to the app to tell it you've done so. When clicking the "done" button, it often takes ~5 minutes to recognize what you've done, sometimes reporting several errors (but still proceeding past them) as it's going. During this time, there's no indication of how long you'll be stuck there, or if the operation will ever actually complete successfully. Sometimes I'd restart the app after a few minutes and try again. This process became a daily occurrence.

It got so bad that I tried resolving conflicts using GitHub's UI, thinking I'd just be able to pull down the changes when done. But GitButler wouldn't have that: it detected the changes and required me to rebase or merge to pull them in. The final straw for me was when both of these options resulted in errors, and I was effectively stuck. Luckily, it's easy to just ignore the workspace branch and switch back to your standard git client.

I think the tool shows a lot of promise, but it's just not ready for use in a real-world working conditions, where you're often integrating your work with that of a team and need to frequently merge it, even in the most common cases where conflicts should be easily resolved. I'll revisit it in a year!

Comments URL: https://news.ycombinator.com/item?id=46767534

Points: 4

# Comments: 1

Rates for my northeast town increased by ~25% in 2024 and are going up by another ~10% this year. It's a hard sell to spend a large amount of up-front money (even after rebates, which decreased this year) to convert to a system that will cost you more than you pay today, and may not work as well in cold weather (every heat pump company I talked to suggested keeping my existing gas heating in place and automatically switching to it when it gets cold enough).

I was also told that the electrical grid in my area is having difficulty keeping up with the push towards heat pumps, which increase load exactly on the coldest nights of the year, when you need heating most.

> If I collect 100 XP, what does it mean for my language skills? For that matter, why do I collect extra XP when I receive a potion? Can the XP I collect be used in a way to carefully guide me towards the specific language skills I would explore next?

Using XP to guide the user towards a particular path is an idea, but it's just not one that Duolingo uses. The purpose of XP in Duolingo is simpler: people like numbers to go up, so they get XP for using the app. It also enables an ecosystem of rewards; I'm generally not a competitive person, and there have still been days where I took a few more Duolingo lessons because I was close to completing a "daily challenge".

Similarly, friend streaks, leaderboards, etc, all have innately appealing hooks. They won't all appeal to everyone all the time, but one of them will appeal to someone some of the time. If they get you to practice for 5m a day more than you would've otherwise, I think they've served their purpose.

Broadly, I agree with other comments about expectation management and time commitment. Could you get yourself to a solid level of understanding in a new language only by using Duolingo? Possibly, but you'd need a lot of dedication and hard work, and much more than 5m a day. If you really wanted to learn a language, and had the time, there are much more effective ways to get there.

Duolingo isn't really built towards encouraging that kind of intense learning, because they know most people who download the app are looking for a bite-sized learning experience, and are willing to accept bite-sized results in return. For myself, I can say that after a couple of years of leaning Spanish on Duolingo, with no previous experience in the language, and an average effort of probably ~10m a day (many days less, some days more), I can read texts if they aren't too complex, follow a casual conversation, and communicate basic things. That's way more than I would've been able to do if I wasn't using the app.

> telling kids the truth before they're ready and without typical parental censorship

Does AI today reliably respond with "the truth"? There are countless documented incidents of even full-grown, extremely well-educated adults (e.g. lawyers) believing well-phased hallucinations. Kids, and particularly small kids who haven't yet had much education about critical thinking and what to believe, have no chance. Conversational AI today isn't an uncensured search engine into a set of well-reasoned facts, it's an algorithm constructing a response based on what it's learned people on the internet want to hear, with no real concept of what's right or wrong, or a foundational set of knowledge about the world to contrast with and validate against.

> what exactly is the fear

Being fed reliable-sounding misinformation is one. Another is being used for emotional support (which kids do even with non-talking stuffed animals), when the AI has no real concept of how to emotionally support a kid and could just as easily do the opposite. I guess overall, the concern is having a kid spend a large amount of time talking to "someone" who sounds very convincing, has no real sense of morality or truth, and can potentially distort their world view in negative ways.

And yea, there's also exposing kids to subjects they're in no way equipped to handle yet, or encouraging them to do something that would result in harm to themselves or to others. Kids are very suggestible, and it takes a long while for them to develop a real understanding of the consequences of their actions.

IMO this is only exacerbated by how little children (who are the presumably the target audience for stuffed animals that talk) often don't follow "normal" patterns of conversation or topics, so it feels like it'd be hard to accurately simulate/test ways in which unexpected & undesirable responses could come out.

Comments URL: https://news.ycombinator.com/item?id=43336548

Points: 2

# Comments: 0

"The cost of treating illnesses has gone up due to X" rather than "more people are getting sicker due to X".

Makes it more of a detached conversation about pricing than about how something is hurting actual people, who perhaps have value beyond a purely economical one.

It's... interesting to see just how easy it is to access this functionality if the VIN check is bypassed.

But my 70yo mother, who is pretty far from being technologically savvy, uses continuity every day to copy one-time-use codes from her phone to her computer, even though she'd have no idea what the term "continuity" means in this context. She notices that it's easier to snap better pictures in more conditions than it was a few years back (and that pictures she receives are better looking on average, too). She uses 1Password with FaceID, which I set up for her, because it's so easy to just look at your phone to unlock that there's very little in the way of enabling and using that, and she doesn't need to write down passwords anymore.

I think some of the magic of the Apple ecosystem is that you don't have to know about these things in order to use them. Someone shows you how to do something (Apple could certainly improve on the organic discoverability of many of these features! Some are impossible to find without looking), and then it often just works. And these things do keep getting closer to that ideal over time, with each generation. When I first started using continuity — long before my mother did — it definitely did not work all the time, and I persisted because I'm a techie early adopter. Eventually, though, it reached a state where once folks learn about it, they can just use it.

I'm also not sure about the 3-4 year number, at least from personal experience, fwiw. We pass down phones in my family, and it easily takes 5-6 generations for them to reach the end of that chain and be in use for a year or two before they're switched out for the next model. Battery has never been the reason someone in that chain switched phones.

> "Ground-Breaking" New Magnet Free of Rare Earth Metals Developed Using Machine Learning

Would've been a more accurate, but less click-baity title, that didn't imply the existence of an artificially intelligent entity developing something by itself. Human researchers used machine learning to comb through many possibilities and find one that fit the bill — which is still interesting and cool.

My first thought after seeing the end of the video was "why is this video flipped?" until I realized that it was only the train that was "upside down".

Comments URL: https://news.ycombinator.com/item?id=40343190

Points: 1

# Comments: 1

It's amazing how even seeing the probabilities written out, or running simulations, doesn't really make it easier to truly understand the result.