I will not be surprised to see less than 20% of all traffic being human. All social media platforms become unbearable at this point. From what I see video hosting like tiktok or insta at least still feels human because requires human face(talking head content) even if they generate scipts with llms they still need to record themselves, that is a bottleneck for army of chatbots taking over discussion. imho of course, and the motivation of particular ppl is completely different discussion.
I genuinely don’t understand how anyone (with technical background) can see LLMs anything more then fancy autocomplete. If you know anything about NNs and about average code quality, that LLMs never will be able to generate high quality code.
Im ready to get downvotes again for my takes, but as a person who writes and trains DL models, I will die on the hill: “people need to produce high quality data” it can be code it can be art, but we can’t rely on those models and trust in the things that they provide.
The new bottle neck isn't writing code, it's testing. You're right you can't blindly trust the output of an LLM but you can trust the testing regime to ensure a certain standard has been met. In hindsight this actually sort of obvious, the more things change the more they stay the same etc.
Well it’s not obvious that is true. If you ask LLM to write tests, it will generate versions of them that code passes, that doesn’t guarantee good code. If you write tests yourself and just pray for great LLM pull, it’s easier to just write code yourself, in my humble opinion
That's a useless approach as you point out but doesn't meant there isn't a valid testing regime to be explored and upheld. Manual testing is going to be a lot more important, I see QA teams/roles becoming very valuable assets in the coming years.
you can trust the tests that you have written, but what about the tests that you didn't write? can you be sure that your testsuite is complete?
when i do test driven development, all the thinking goes into the tests, and the actual code writes itself. LLMs hardly help make that any faster.
having a complete testsuite may make it easier to use LLMs for refactoring, and adding features, but then you still have to write he tests for the new functionality.
Because often the problems that these people are working on tend to be trivial. LLMs are excellent for making the millionth CRUD backend server that talks to SQLite and a glorified todo list React frontend. In fact, it's stupid to do it any other way now.
Obviously I familiar with RL, written multiple training pipelines in my day. and in order to gain that “super human skill” using RL you need to define fit functions and provide environments that will provide you with feedback that used for training. Go and chess are have clear rules and environment that provide you with a signal of success, I waiting to see this for coding, I don’t say it’s impossible just orders of magnitude harder
when i was in uni in 2014 i learned that smalltalk became obsolete, later i went to industry to see that no one use smalltalk(or prolog) and yet on this site ppl bringing up smalltalk every single month, why is that i wonder
Yeah there's a difference between being popular and influential. Smalltalk is influential without being all that popular.
I suspect that's largely because most smalltalk implementations don't store code in git. "A smalltalk" is more like a Linux distro than a phone app. It's almost like the main purpose of a smalltalk is to build more smalltalks. It's a good way for developers to build a deep, powerful tool for other developers, but it's a less good way to build user experiences targeted at non-programmers, who are going to want to have a bunch of different apps to use not a bunch of different operating systems to use.
Good economics for the classroom; bad economics for the real world. Then and now, though, a great way to play with novel ideas in programming languages.
The Potsdam university (near Berlin, Germany) and Hasno Platner Instute [1] has been actively teaching and researching Squeak Smalltalk for decades. Same in Buenos Aires and several other places. Science papers every month for 5 decades, under many names besides Smalltalk. Weekly online conferences, presentations.
People like the language - I like it though I have only written a few bits and bobs in it. They also like Fortran, Prolog, and (dare I say it) Haskell, and many others that don't appear here so often. Just because a language is not this week's hotness (looking at you Rust) does not make it "obsolete".
Likely because whilst it didn't work out commercially, the ideas smalltalk, prolog and other more esoteric languages (forth, apl) focus on are themselves very interesting.
In what way, obsolete? There is quite a bit of worthwhile research done in Smalltalk-related languages and some corporations are quite successful in building on such languages.
It's not dominant in the way of Wordpress but also not insignificant.
A lot of systems are obsolete, in the sense that you cannot find work in them. Nonetheless, people use them to prototype ideas that will eventually find their way into industry. Is Haskell obsolete? Not exactly, but it's hard to find real-world projects that use it. Concepts developed in Haskell get imported into other systems.
Systems are only truly obsolete when there is no real reason to use them. PL/1 is obsolete. It's hard to find even legacy code using it, and nobody would use it for experimentation.
mostly because everything relevant we have today was built on smalltalk ideas and it still executed them better in the 70s than modern languages do today.
