I don't get this sentiment. Think of the future innovations possible with faster research and computation - space exploration, DNA-based health improvements, robotic helpers - read a few sci-fi books to imagine what the future can be! Computer science doesn't have to end with everybody getting laid off due to no more CRUD apps needed.
I find it's easy to scroll past very sloppy posts. And 1 in 2 posts are genuinely interesting/teach me something new about the world and the people in it - how people conduct business - and most of all their improbable trajectories that got them where they are currently in their career. It's very interesting and the recommendations are great for me. I don't find too much slop or cringe content, and am confused when people complain about it so much. It feels like people don't really know how to separate slop content (self-promotion/empty AI posts) from genuinely written life and work experiences. And HN is definitely an echo-chamber, I see a huge amount of people (more than instagram tbh in mid-30s age bracket) contribute regularly to linkedin.
Am I too pessimistic, or did anyone else notice the fact that you got everything you wanted with 'the only other asian guy'. Race matters in friendships more than we realize? I feel if it was a gym full of asian people (or a neighborhood with more asians) it would be way easier for you to socialize and make new friends.
Tried generating a 6-month plan to upskill in VLSI programming. It was pretty useful and helped lay out a comprehensive roadmap using different sources and projects to try.
Also tried generating a 6-month plan to upskill my guitar playing. Wasn't as useful.
Musk has basically discovered that you can ignore existing laws, since by the time lawyers sue and courts order injunctions, it'll be too late and too expensive. Especially when lawyers can argue against basic facts like "Musk doesn't head DOGE". It's the same playbook as the twitter layoffs - when you are so rich, you don't need to care about laws.
30000 hours seems like a lot, more than enough time for anything you might want to do. More than that, there might be enough changes in technology and the world that something you learn in the first 100 hours might not matter in the last 100 hours etc. If you start taking all these complexities into account, it might be better to count at a granularity of weeks rather than hours. Or maybe even ignore time, and focus on goals and aims. You can always make more or less time to fit your goals.
Another complication - H1Bs are valid for 6 years. However, the physical paper stamp given on passports is only valid for 3 years at a time. So if I chose to stay in the US (and not travel internationally) I'd be fine to continue working without renewing my physical visa paper stamp. But to get a new physical stamp, I have to re-enter the country every 3 years. How does that make sense?
How about Discord? Or reddit? Would those be limited to kids as well? You can be sure that even if FB/Instagram were to ban kids, there would be hundreds of companies jumping in to scoop up all that teenage DAU.
Even if Facebook implements these controls, what guarantee do we have that another social app won't come along without these controls? Do we regulate TikTok, Youtube, Snapchat, GroupMe or whatever the latest flavor of the month is as well? There are probably thousands of startups that would jump at the chance to monetize teenagers even if FB were to step aside.
It's especially bad because I don't really remember my passwords, so I always have to reset password when logging in again, and that refreshes the dev keys so my terminal git push also stops works - a complete PITA.
It's like saying, I hope the new pop music trends that all the kids are listening to dies out, and 80s rock becomes cool again. It just isn't gonna happen.
When you click a tweet that is part of a thread and is several tweets down from the top, it goes into that tweet's page. Then when you press the browser back button, it should (and used to) save the scroll position of the tweet you were last at. But now it always resets you back to the top starting of the thread. Super irritating when you want to read replies for specific tweets in a tweetstorm.
The first use case I could think of was a AI trained on my team's codebase. So that whenever we face an oncall incident or unexpected crash, we'd have this AI partner (that has been finetuned on all of our code / logs / previous bugs / documentation) that we could ask to help us debug.
Parent was definitely a chatGPT comment. The 'However' in the 2nd line, the neutral and unbalanced tone, and overall the informative and well constructed paragraph without any grammatical errors or personal ego - you'd identify it immediately if you played with chatGPT for a while.
I get a feeling that the twitter mob/vigilante journalism unleashed upon the FTX meltdown story may not entirely be accurate. Obviously something bad has happened financially, and people and emotional and angry. But we have to give our courts and the law system a chance. I don't want vigilante mobs condemning people to jail and death. We are literally hearing only one side of the story - maybe there is a 1% chance that the books were not cooked, that the rapid crypto crash combined with bad accounting practices (not fraud) did actually lead to losing billions of dollars. The very fact that SBF/caroline haven't been arrested leads me to think that the legal authorities see facts which we cannot see. There is an echo chamber forming that is
- republican (people linking this meltdown to deep state conspiracy theories)
- sexist (comments about caroline)
- anti-semitic (comments about religion)
and in the past, these mobs have been wrong in many cases. People are repeating theories as if they are facts, and that makes me uneasy.
Probably because there's a lot more material and detail about computer science online that it would be regurgitating. You can be sure that someone has already answered whatever CS-related question you are asking it verbatim - all it does is connect the dots and glue the concepts together.
Perhaps in the future the company would not be adding any new models or require optimization of any new surfaces, so they don't expect to be spending enough on new initiatives to justify optimizing them. And all the existing initiatives have been optimized efficiently already (though that does seem unlikely when I type it out).
Start a part-time masters online or at a local uni. You'll meet plenty of motivated people. Btw, you might find that most people are motivated by money, not by the love of science/tinkering.
I've heard of some of the famous viruses which are obviously more complicated than what I mentioned. But that's the difference - they are hardcoded to perform these particular steps. I'm not saying AGI is the code that does this. AGI is one level higher - a more meta objective - given the objective of "accumulating power", it will learn lower level objectives like "replicating and spreading", which leads to the final output like this code.
Basically, AGI could unleash a thousand "morris worms" - and that would only be for the objective of replicating and spreading. There could be other objectives as well.