Is it possible to add some kind of self evaluation to the answers given by a model? Like, how confident is it with its answers.
HN user
sboomer
I don't know much about language models, but don't they just have an understanding/knowledge of patterns between words, and don't have the reasoning capability at all?
Some content is only for premium subscribers. When you try to log in/subscribe the site throws php exception with error stack as if it is running on dev environment.
I have used this to solve leetcode problems. Initially, I read and understand the problem; and then I try to solve it for ten minutes. If fail to find a solution, I move on to the next problem; I do that for 3–4 problems on a single go. When I visit those problems the next day, I find myself hitting the solution way more than I would be if I had sat there for hours with hard focus to solve those problems. However, I feel that I cheated because you are expected to solve those problems in limited time.
I experience a similar kind of state before I fall asleep, but not every day. I sense the trigger of going into it, but most of the time I suppress it due to tiredness or lack of interest. However, there are certain days during which, I go into the state of super alertness, where I get amazing clarity of thoughts; one would wish to be in it while solving a new coding problem or trying to write super intricate code.
There are many people who haven't seen the internet before 2010. Their internet experience of limited to Facebook, Instagram, YouTube, and WhatsApp.
I believe Discord and Telegram may fill in the gap. WhatsApp is kind of personal, where you communicate with your real world friends.
It can bring with it the anxiety of living up to their expectations. This is present for big celebs too, but in this case the pressure is amplified because you are kind of close to these 1000 or so guys, and they pay you.
I feel like most people are trying to give personality to AI chatbots. From what I have read, ChatGPT is good at carrying conversations and doing text based work. I would focus on feeding it curated info and making its answers as factually correct as possible. And some way to know the sources of information it picked up answers from. Companies and people are going to do that anyway in the coming future. We may see chatbots good at a certain field, and whose results can be trusted.
With or without caching?
It may work if any of the mainstream Linux distros run on your pc.
I would have loved to have this as my secondary os. It's shame that one cannot multi boot this. I'm looking for a way to multi boot this.
That's how a software implementation by a newbie works. You can't expect a newbie to take security into account before the software is implemented. Instead, there should be a custom to rectify all the security errors in the end before the software is pushed to the server.
It would be nice to give an option to download an image of this page in required dimensions. I would really love to have these as my wallpaper.
Roaming around the city and visiting places with my friends.
Being cared and loved by the person you like.
The feeling of contentment before sleep when I was productive on that day.
You can go and live in the woods enjoying your solitude, but there is always that part in your heart that needs to be filled with being around loving human beings.
Any millennial who is using search for some time would easily know where to find what he needs. This sounds like Google is trying hard to drive more money out of its search business.