> I regularly see people claim that they are now 10x more productive with LLM code generation, and I just wonder where all the code is.
Perhaps the biggest issue hasn't been the speed of development all along but people working on wrong things, hence why progress isn't as evident. LLM helping to build something that isn't needed faster isn't going to productivity more visible.
There are already plenty of SaaS that offer open source version as a way to push to the cloud and charge for hosting, maintenance, support and some extra features. In that way, a working code have already been commoditized and those companies didn't disappear.
This. First I try it just a little to do a boring part. It feels great. The boring part that was holding me is gone and all it took was a little instruction. The dopamine hit is real. So of course I will try it again. But not so fast. It needs to be corrected to make everything aligned with the architecture. And as my requests get bigger, it needs more and more corrections. Eventually correcting everything becomes too tedious, and accepting is just too easy, and so I lower my standards, and soon enough lose track of all the decisions. The branch is now useless as I don't want to debug or own this code I no longer understand hence I start over. I want work to felt like a training session where you get fairly rewarded for your efforts with better understanding, not like a slot machine where you passively hope it gets it right next time.
I had to use Borland when studying but never really liked it. Later I tried Notepad+ with console for the first time and everything clicked. There is a console for compiling and running and there is a text editor for editing. It was the flow that made sense.
Database is one of those places where it's justified, I think. Application containers do not need the same level of care hence are easy to run yourself.
Not sure about windows but I solved it for myself with basic provisioning script (could be an ansible playbook also) that installs everything on a fresh linux vm in a few minutes. For macos, there is tart vm that works well with arm64 (very little overhead compared to alternatives). Could be a rented cloud vm in a nearby location with low latency. Being a neovim user also helped not to having to worry about file sync when editing.
Sounds like social coaching has all the attributes of an untapped business niche. The question is, if there is just a obvious pain, why there is such a shortage of solutions. I think it could be because of the stigma associated with this issue or because people affect by it lack genuine interest in solutions as people are usually happy to pay for all kinds of self-improvements.
I think it needs to know every conversation that happened in the company to provide meaningful feedback that is above generic coding suggestions. But then it makes onboarding new people and getting them familiar with the project much easier. Businesses will have a lot of incentives to do that.
So they lie about estimates because they afraid to miss their chance to delivery buggy mess. But doesn't it do more harm to the reputation in the long term?
Trying Gleam that also runs on Erlang VM but has static types, a fast compiler with Rust-style elaborate error messages made me hopeful about how much nicer this experience can be when static types are added to Elixir too.
Exactly. It doesn't need extra buttons in the UI, it needs a giant context window to contain all the important info about your life or a business and the tools for gathering and maintaining that context. Then merely saying that you are seek becomes enough to autocomplete the rest of the actions including sending an email.
This is my thinking too, yet for some reason, 90% of job desciptions insist the candidate has N years of experience with M stack. As if someone who was building websites with Rails for 10 years is somehow not qualified for building the same websites with Symphony.
There should be a standard way for all browsers to tell sites whether they accept cookies or not. There is no reason every user should deal with this banner manually for all sites.
I think it's possible that in the end, AI will make everyone wealthier nevertheless. Just like people today posses the level of conveniences unimaginable to the elites of the past, in the form of smartphones, global delivery, cheap flights, instant access to information, etc. Being able to afford an unlimited, available 24/7 health-related consultation for $20/mo is also wealth and so is being able to single-handedly create an app that would otherwise require a team of 10.
Also, it seems to be that information that is helpful just doesn't like to be contained. Comparing with StackOverflow, its popularity didn't make developers less likely to participate in the community. Instead it made programming more approachable to a much larger pool of people and more software were created, which made our life easier. If something is intended to be used only for consumption (media) it tend to say closed. But if something can become a building block for others, people generally seem to want it to spread.
As someone who considered giving up on programming due to wrist pain from typing, this is my dream as well. I also prefer to work independently or in small teams, and the amount of things that could be tackled this way is surely increasing. There is no reason why programming shouldn't to be more about thinking and less about typing.
I see limited context and the lack of sandbox dev environment is the main obstacle on way to have more sophisticated usage and faster feedback loop. The value would be greater if it could know everything about the code base, business requirements, related documentation, and, it could even avoid a lot mistakes by simple running type checks and the tests and correct itself before giving the answer.
They do help with promotion by giving "top rated" and "raising talent" badges and work history. Hard to say how much value and leverage it really gives and whether it worths all the downsides.
I wonder if the lack of libraries in a new language is just a few ChatGPT-integrations away. After all, Crystal is so similar to Ruby. I bet as soon as context/environment aware ChatGPT tools/agents that can compile / run tests / apply fixes until it works are available, this will become the reality: "Create this Ruby library but for Crystal" and in a few hours you can include it in your Crystal project.
Writing code is one thing, owning it and having responsibility is another. But can we build tooling for this to on top of LLMs too? This won't happen overnight but it will certainly happen gradually, I don't see why not. By owning it, I mean, upgrading libraries, fixing bugs, writing tests, running tests, fixing tests, and maybe even adding features.
It's totally possible to make LLMs write actually good code that can be maintained using normal software engineering practices.
Of course there will still be a need for a human touch, such as knowing limitations of LLMs and possible workarounds and understanding business requirements. Such productivity boost may even create more jobs as products will become cheaper and faster to create, hence more new markets will be explored. The pie might also get bigger not smaller.
Travelling changed the perspective so much of me, I realised that a lot of things I was hesitant to do or even thinking was directed by the unpalatable attachment to the static environment.
Imagine the world where everything is fully automated. You arrive at a city, order a self-driving taxi, do a self-check-in at the hotel that is managed by robots, order a food that is cooked by robots, using ingredients delivered by robots, grown in farms managed by robots. The hotel itself was 3d printed and the city district was planned by an AI algorithm optimised for economic utility. How many iterations are there before this is the reality? It would probably take a century before something like this is the norm.
It's the power that is scary, the power that you cannot control. It changes the balance completely. With AGI, humans are no longer in charge, therefore are irrelevant.