759 karma · joined November 3, 2015
It's quite possible there is some lower altitude fog/precipitation/something that is only visible to the radar at it's lower tilt angles. But that's just speculation.
I think this is partially why people want these types of objects. You can actually own them. You can understand them. You can mod them.
Modern cars, games, computers, etc. aren't owned anymore. They always have updates, and could be bricked at any moment if the manufacturer wishes, gets bought or goes out of business.
There's nothing preventing this from being the golden age of computers. The capabilities of the hardware are near magical. We just need to bring back the concept of ownership.
I'm really hoping raspberry pis can fill the roll of a "standard computer". Currently, they have pretty rapidly changing CPUs, etc. I would like to see a pi that settles down as "The Pi" that doesn't change for decades. This would allow all kinds of software (and OSs) to be written for it that stands the test of time. Download a binary and it just works. No need for updates/changes etc.
I've found the documentation for Janet to be fairly adequate. What are some parts where the documentation could be better?
I'm going to laugh pretty hard if it could just be done with: cat file1.csv file2.csv > combined.csv
Proper nouns are trivia though. The more important part of the document is that students couldn't tell the difference between what was literal and what was metaphor. For students that were in their 3rd year of university as English majors, that's quite a problem. By that point they should have read hundreds of passages from books written in the 19th and 20th centuries.
I've recently started playing around with Janet, and it's a great language. I think it's inspired by Clojure and Lua, and somehow manages to be better than both (in my opinion).
Also I think Odin struck the right balance of features while trying to keep to the spirit of C.
Also, decent human beings care about things that are unjust/immoral/unethical regardless if it affects them or not.
I think it's a better idea to write a function that applies a function to the elements of a tree. Ideally you'd make a function for each traversal order. This makes it obvious what is happening.
map_tree_pre(my_tree, my_func);
map_tree_in(my_tree, my_func);
map_tree_post(my_tree, my_func);
map_tree_bfs(my_tree, my_func);
It reminds me of an experience I had at work. I wrote some tests for a feature that would store some values in a database. However, my tests would always fail. Retrieved data would never match the input data. I eventually decided to trace exactly what was happening in the database and found a stored procedure ran whenever I entered data. It "fixed" my input data. When I told another dev about this they said their "database expert" said it was "good design".
Absolute insanity. Never, ever, "correct" deliberate input. Validate it and tell the user what is wrong. 1) This raises awareness 2) The desired behavior is easier to test.
I use AI rarely, but I use it to try to learn. For example, gnuplot is pretty complicated, so I'll ask it to create some examples and that's allowed me to learn way faster than trying to parse out the man page. AI is great for this, even though that's not how most people use it.
Some parts of software dev will likely be automated away (just like compilers automated conversion to binary and some parts of optimization), but I think ultimately software dev will become more important over time. Natural languages have low information density and high ambiguity. Programming languages are the opposite. So I think programming will always be a thing.
Think about music. It's not written in natural language, even though it could. We continue to use musical notation because it is the best way to communicate music. I believe the best way to have unambiguous, information-dense, logical communication with a computer is with a programming language.
I will bet this CEO $1 billion that one of his employees using AI for 3 days will not be able to out-produce me working for a year.
Either what we have now is important and will be used 20 years from now and you should learn it, or things are improving so fast that our future interface will be very different.
How many of the "prompt-engineering" tricks that people gained from 2 years ago are relevant today? Almost none? So why should anyone waste their time on something that isn't very good today?
Use LLMs. Try them out for sure. But don't buy into the VC hype. They're just trying to 10x their money by spreading FUD.
This is the talk that introduced Upspin: https://www.youtube.com/watch?v=ENLWEfi0Tkg