I would not want to be a “take this UI and add a new field” coder right now.
First, there's the uselessness of entry-level engineers. They are hired for their growth potential (that every human has, but language models don't) and are expected to grow into mid-levels and seniors. In most of the sw industry, entry-level roles are also not terminal, which means that such an engineer must get promoted at least +1L to keep working in the company. Someone (or something) perpetually stuck at the entry level is a bad value proposition for a sw company.
Secondly, the code that language models produce is buggy. They can, on occasion, produce amazing code and even entire codebases. But this is an exception, not the rule. You can generally prototype something or get an idea for something from language models, but you can only push very little of that into production. What good is code you can't use? You still need an engineer to oversee the language model's outputs.
Overall, if some company replaced their juniors with AI, that would be incompetent management.
We've never had glorified chatbots like GPT-3 or GPT-3.5. I'm not just praising GPT; I've myself casually run through a few simulations with a hotel chain receptionist and an executive. The technology looks very competent, and aside from cost savings, there's also the customer service quality aspect (consistency particularly) and the element of removing staff from abuse.
The biggest challenge is integrating language models with live data. Making customer data accessible to them is not a problem technically (prepending prompts), but it could be a GDPR problem if a third party like OpenAI is involved (having to hand over data to a third party might make the AI receptionists unappealing to some customers -- maybe, needs to be tried). The other aspect is letting the AI make changes in a data model. But there are ways to solve that as well. When these obstacles are resolved - and there is a lot of incentive to fix them now - a lot of customer-facing reception-type work can be outsourced to AI.
By the way, some hotels are already very interested in chatbots for reception work. There has been a lot of talk about that in some chains since about 2020. Old-style NLP bots, too. But of course, GPT-3 capabilities are very appealing.
Kinda amazing this has not caught on more and sooner in USA.
Much more low tech versions of this have been used in Japan forever. The noodle bars and ramen shops are case studies in efficiency.
Of course. My comment was about 'less programmers' getting hired. Not the extreme of 'All programmers' getting replaced by ChatGPT as you read incorrectly.
Believing that programmers don't need to adapt and that managers won't need to reduce or do layoffs especially with programmers, is quite frankly a denial and delusion of what is going on with cutting costs in hiring too many of them in the first place.
Denial is the first step to acceptance and as I said before, less programmers will be hired with juniors and some seniors impacted.
More developers will get hired, because ChatGPT will not perform the higher function tasks most developers actually spend most of their time working on.
The reason this will happen, is because ChatGPT will be more efficient in performing the lower function tasks and there will be more of a need to put all that automatically created code into actual use, integrate it with Enterprise Package FooCPT, figure out what caused the product outage at 2am at night. And on and on and on.
Not to mention that ChatGPT will not be able to talk to your users about what they want next, convince your manager that that's the right thing to do and everything else that you need to do on top of the coding part to be a good developer.
What possibly will happen is that junior developers will have even harder time getting into the industry, because clueless PHBs will see ChatGPT doing all their work "for free". A few years later they're screwed, because nobody at the company knows how their applications work. Except ChatGPT...maybe.
Reads like what I said before. Some will be affected and less programmers will be needed,. Especially in particular, juniors and some seniors getting impacted.
I don't know where I said 'All jobs' as you once again brought out as you clear read incorrectly. I only said less programmers will be hired and these software engineering jobs will be more competitive.
The best course of action is to learn to adapt and it certainly applies to programmers. Pretending that they don't need to adapt is a delusion and denial of constant change. With or without AI.
The performance gains from chatgpt are going to be amazing from a quality of life perspective but overall not a huge change (except maybe in some niche roles). Programmers are not spending much of their time figuring out code that chatgpt can generate.
You are like a peasant arguing about security of their job in the beginning of 20th century after seeing a tractor. Sure there are farmers now, but much fewer of them is needed.
To anyone who used Copilot and ChatGPT it is obvious that in 5 years demands for entry to software dev will be much higher, and there will be less positions open, because sheer productivity increases for the already employed people will outpace the demand. It’s like peasants arguing about security of their future prospects in the beginning of 20th century after seeing a tractor.
Winter is coming, but they just can’t believe this since their income directly depends on this not being true.
Read a freaking newspaper once in a while.
But of course. Denying it all, and not adapting sure is a great way of coping with rapid changes. /s
We’ll get another slew of decades long court cases once one line of Oracle or Microsoft code proprietary code comes out of chatGPT.
So everyone who knows how to do a for loop is safe.
Or the extreme, we finally have to move to post-money society, because automation is no longer creating any jobs.
I still stand by that no-one (including Big Tech companies) is safe, especially after the 2022 tech layoffs which happened to many and tools like ChatGPT will eventually reduce the need to hire lots of programmers like before.
It is not going to be the same as the hiring euphoria that happened last year.
So, you have a choice how to look at the world.
I don't understand the relevance of ChatGPT to this.
- write a Gitlab ci job that tests python code and outputs a junit test report.
- can you add code coverage?
- write a sql table that has users, age, and a biography
- write python that parses the above output from the table
- now use a lexer and parser
- create a unit test to verify it's correct.
- write a perl one liner that parses a list of users and phone numbers
It completed all of them. I was already impressed with OpenAI when it generated an argument for why Britney Spears should become president and a test for a third grader about Socrates which I then answered incorrectly and had it grade it.
It is often confidently incorrect. Which is easy to spot if you know the subject matter, but nearly impossible if you are not familiar with the subject.
My god, it truly can replace computer programmers.
I found it funny because it made it more human-like as well.
(Example: “what’s the difference between timestamp and timestamptz in postgresql?” will answer that timestamptz will store the time zone and takes more space, which is both incorrect!)
I also asked it general, abstract programming advice and it gave pretty well-reasoned arguments mentioning maintainability and readability.
Garbage in, Garbage out.
GPT3 is as good and reliable as a Tesla autopilot.
Why you ask? Well it's because most of what a programmer does is more complex than the AI can handle right now. For example my job involves the following:
a) multiple interacting systems and management of the relationships between them
b) > 90% of my work does not involve writing new code
c) the programming part is mostly about changing existing code
d) or debugging (and sometimes setting up a debugger and running code)
e) reading hundreds of articles to figure out why the trivial example these systems are good at doesn't work (sometimes on someone else's machine or staging only)
f) taking product requirements and making small alterations to allow them to become reality (and knowing when to tell people about the decisions you made and when not to)
g) talking to people who might know more about the specific problem you are facing and when to battle through
h) sharing ideas on how to do things with the wider team
i) fighting the good fight and sneaking in refactoring when people are not looking
I don't think we are yet anywhere near these things yet.
Perhaps now one may say, it is time to 'learn to adapt', since with the recent release of ChatGPT now gives the excuse for managers to be more motivated to hire less engineers and programmers as the cheap money has essentially evaporated.
So this time it is indeed different and I would expect less engineers to be hired and more programming jobs to become very competitive. No-one is safe; and that includes 'seniors'.