- We need some regulation on them for sure. They should be paying for the content they train on and use in their search results.
They’re still very compelling as a user.
- I totally agree. It’s just going to become an expectation that AI is in the browser.
It’s so nice just to be able to ask the browser to summarize the page, or ask questions about a long article.
I know a lot of people on Hacker News are hostile to AI and like to imagine everybody hates it, but I personally find it very helpful.
- Typically I think, but you could pre-train your previous model on new data too.
I don’t think it’s publicly known for sure how different the models really are. You can improve a lot just by improving the post-training set.
- There have been bugs and regressions since forever. It’s easy to look back with rose colored glasses, but I don’t think software has actually gotten worse.
Just look back at the Snow Leopard release of OS X. It was specifically marketed at having no new features and just being a fix and optimization release because Leopard was such a mess. And people were happy about this.
- If you are reducing the likelihood of something by 99%, you are basically eliminating it. Not fully, but it’s still a huge improvement.
It reminds me of this fun question:
What’s the difference between a million dollars and a billion dollars? A billion dollars.
A million dollars is a lot of money to most people, but it’s effectively nothing compared to a billion dollars.
- I’m not so sure. The big selling point for Rust is making memory management safe without significant overhead.
Zig, for all its ergonomic benefits, doesn’t make memory management safe like Rust does.
I kind of doubt the Linux maintainers would want to introduce a third language to the codebase.
And it seems unlikely they’d go through all the effort of porting safer Rust code into less safe Zig code just for ergonomics.
- What other newfangled alternative to C was ever adopted in the Linux kernel?
I have no doubt C will be around for a long time, but I think Rust also has a lot of staying power and won’t soon be replaced.
- There actually are techniques to win (and presumably lose) rock paper scissors more frequently than random due to psychological factors that make player choices somewhat predictable.
- They do have user generated lists to help find things to do, but yeah, I mainly use it for organizing ideas and building my final itinerary. It’s really nice for this. I can then also share the itinerary with everybody else on the trip.
- “They’ll have to” is a pretty strong statement assuming you don’t actually have insider information on their API spending.
For now, they do use Google Maps and I’m happy with it. If they stop and it’s no longer as useful to me, maybe I’ll stop using it.
- I use and pay for Wanderlog. Idk how their business is doing, but I love it as a user. They use an embedded Google Maps viewer for locations, so there is no problem for coverage.
- Waymo continues to improve every year, but dumb drivers never will.
- Discord, VS Code, and Figma are all apps that individuals choose and are well liked despite many alternatives. Slack too I think, though I don’t have experience with it.
Your comment applies to Teams and I’m sure other electron apps. But the sweeping generalization that electron apps have terrible user experiences is pretty obviously incorrect.
- They work great for me.
- Not really. The downsides are mostly overblown.
Plenty of category leading applications like Discord, VSCode, Slack, Figma, etc. use it quite successfully.
- As wrong as it feels to have to use Electron for a desktop app, it really is the safest approach for most applications.
Qt also seems to be a good option, though there are licensing considerations for commercial applications.
I’m excited for various upcoming Rust options as well, but right now Electron is the battle tested option.
I am curious though about Avalonia. I’ve heard good things, but it’s definitely a smaller player compared to Electron. I’d most likely choose it over Microsoft’s first party frameworks.
- What you call whataboutism, I call comparative evaluation.
On an absolute scale, every world power in history and humanity in general has been terrible. Endless injustices and atrocities since the very start.
That’s just not a very useful way to look at things though.
- The US hasn't been perfect, but you can hardly say we've ruled the world terribly. Because who has ever ruled it better?
We helped Europe and Asia rebuild after WWII instead of conquering them. To the extent that our previous enemies in Germany and Japan now have some of the strongest economies in the world.
There have certainly been wars, often with dubious justification or horrific results, but good luck finding any superpower in history that hasn't gotten into bad wars. Unlike the US, most of the time those other superpowers used war for territorial expansion, like Russia is doing in Ukraine today.
You can dream of your utopian world order all you want, but at some point you have to judge the US against the alternative instead of the almighty.
I just think the average browser user in 5-10 years will expect the AI features. And plenty of others won’t want to use those features, and that’s fine.