4 ms·

If you want a long lasting career in programming, work on building the languages, not just using the languages. Is probably what someone said many decades ago.

Terr_ 3y ago
And in those decades they were probably right.
xmprt 3y ago
Is that not true to some extent? In the early days of computer science, when there were so many new developments, it would serve you well to be the person working on writing the programs and building computer hardware, rather than the person using computers to get stuff done. The person using computers was still miles ahead of the person not using computers but where you really wanted to be was building languages and inventing new tools.
- sshine 3y ago
  > where you really wanted to be was building languages and inventing new tools. Building computers, operating systems and compilers was always a niche. So if you're a computer scientist or an electrical engineer, sure. But if you're not, there was, is and will be "power users" who make the most of it. - Teach, fix shit and bridge the gap for mortals - Enjoy technology with a deep but not foundational understanding.
parpfish 3y ago
learning a programming language confers useful skills that you can generalize to new versions of that language or different languages. we can make these generalization becasue we know how the langauges work and people can draw connections from one to the other (i think people here would all agree that learning your second language is far easier than leraning your first). this is different from LLMs because they don't have explicitly constructed abilities that we can compare across models. so we need to step back and approach every model as brand new and figure out what they're capable of. just because you have a prompt that works amazingly in one model, there's no guarantee that that prompt will continue to work in bigger "better" models. so rather than working to devise way to trick a particular model into doing your task, it would probably be a better use of your time to learn how to train/modify models to explicitly solve the problem you care about.

Terr_ 3y ago

And in those decades they were probably right.

xmprt 3y ago

Is that not true to some extent? In the early days of computer science, when there were so many new developments, it would serve you well to be the person working on writing the programs and building computer hardware, rather than the person using computers to get stuff done. The person using computers was still miles ahead of the person not using computers but where you really wanted to be was building languages and inventing new tools.

sshine 3y ago
> where you really wanted to be was building languages and inventing new tools. Building computers, operating systems and compilers was always a niche. So if you're a computer scientist or an electrical engineer, sure. But if you're not, there was, is and will be "power users" who make the most of it. - Teach, fix shit and bridge the gap for mortals - Enjoy technology with a deep but not foundational understanding.

parpfish 3y ago

learning a programming language confers useful skills that you can generalize to new versions of that language or different languages. we can make these generalization becasue we know how the langauges work and people can draw connections from one to the other (i think people here would all agree that learning your second language is far easier than leraning your first). this is different from LLMs because they don't have explicitly constructed abilities that we can compare across models. so we need to step back and approach every model as brand new and figure out what they're capable of. just because you have a prompt that works amazingly in one model, there's no guarantee that that prompt will continue to work in bigger "better" models. so rather than working to devise way to trick a particular model into doing your task, it would probably be a better use of your time to learn how to train/modify models to explicitly solve the problem you care about.