barefeg

I have been thinking of something a bit more on the middle. Since there are already useful service APIs, I would first try the following:

1. Describe a set of “tasks” (which map to APIs) and have GPT choose the ones it thinks will solve the user request.

2. Describe to GPT the parameters of each of the selected tasks, and have it choose the values.

3. (Optional) allow GPT to transform the results (assuming all the APIs use the same serialization)

4. Render the response in a frontend and allow the user to give further instructions.

5. Go to 1 but now taking into account the context of the previous response

Off topic and probably the wrong place to ask, but how are the posts ranked? I have also seen posts that are very high but with little to no comments

I have a similar experience to yours; programmed from a young age, did several personal and freelance projects, etc.

It was hard for people to even consider my CV until I shifted focus towards those projects and freelance jobs. Still, the job search wasn’t successful until one person decided to “just give me a chance”. His point of view, which I completely agree with, was that even with a lot of programming experience, he can’t be sure that I’d like working in the software industry. Fortunately I didn’t mind the process and the rest is history.

If I had to do it over again I would first figure out how software engineers really do work, that is learn the different team configurations, how engineers collaborate with each other and product people, the different frameworks used (like agile, XP, plan-build-release, etc.). Then if you decide you still like it, either try to connect your past work experience with this way of working, or try to acquire such experience as others have recommended (contractor work, open source contributor, consultancy work, etc.).

The way I see it is that Google (search engines) is a particular solution to a set of retrieval information problems. ChatGPT does seem to be a good solution for a different set of problems. As others have pointed out both have their benefits and limitations.

The future of how we retrieve knowledge might use a combination of these technologies along with others that may not yet exist.

ChatGPT is definitely showcasing how good the future systems will be at understanding natural language input, as well as producing answers in the appropriate form.

My mobile provider charges a ridiculous amount per mb while the price in a subscription is much cheaper, if you don’t use more nor less of the allocated amount. So even though they provide both options you mentioned, neither feels ethical to me.

Edit: 1 eur per 10mb in the prepaid version vs 13.50 eur per month for 2gb plus other stuff like calls and sms. So I either have to waste money on stuff I don’t use every month or pay 15 times more for the same service

I’m wondering if replicating synchronous communication more closely could help. For example making it not possible to send you a message when you are “in conversation” with another user.

Could you elaborate? I’m in the AI field myself but as a software engineer. But have a PhD in physics

Kind of topic. But I’m wondering if it’s possible to get comparable pricing when self-hosting a solution similar to this in the cloud vs say AWS S3.

Makes sense. I’m not very experienced in privacy but could you explain why uniquely identifying the user is a problem? As in you can tell that there’s one user who visited 100 times but how can you use that information to correlate with an identity?

For neovim I definitely experienced input lag. In VSCode is a bit more subtle; for example when opening new files and saving files with autoformatter

I recommend any of the talks by James Powell at PyData. For example this one https://youtu.be/cKPlPJyQrt4

Edit: maybe this one on Numpy may be more relevant: https://youtu.be/u2yvNw49AX4

I like the progress that is being made for running containerized workloads on macOS. In my case I like some of the benefits of running the workload on a remote machine; such as no fan noise, less heat, less power consumption (especially on laptops). However the downsides can be also quite annoying, such as file sync times or IDE input lag.

My current setup is to have both data and workload run on a remote machine and I connect to it via SSH. I can either run neovim inside or use the remote development plugin from VSCode. But as mentioned, the input lag can be very annoying. I’m wondering if there’s another setup where I can still retain some of the upsides of running the workloads remotely and still having a decent user experience (reduced lag)

I love neovim but I don’t have a good replacement for the git plugin of vscode. I haven’t found my way to replicating that experience in neovim. I tried lazygit but I find the diff viewer hard to understand and navigate. Diff view (https://github.com/sindrets/diffview.nvim) might be it, but I haven’t made the time to install a newer version of git from source (Debian…)

I recently got into neovim (from vscode) and it took me 1.5 weeks to get it to have similar functionality. I really like the vim commands though. What’s your opinion on that compared to helix?

Ah that’s a great suggestion! I didn’t think about it even though I use kustomize a lot.

About the namespace, I usually modify it manually if it’s a few files, otherwise use some post processing like https://github.com/helm/helm/issues/3553#issuecomment-417800...

> My favoured method is to keep manifests in full (which you can source from `helm template`!) as pure yaml files and version that.

I do something like this but I normally find that helm charts are not parametrized as I want and have to manually modify the output manifests. When updating from helm it can be challenging for other team members to understand what bits we want to take from the new helm template output and which ones we don’t. How do you deal with this?

Sometimes I update the helm chart to fit our use case, but it’s still hard if that is not merged upstream (because that means maintaining our own version of the helm chart)

Sorry for the unfamiliarity with JetBrains products. Is this architecture similar to VSCode’s? So the ide has a backend that can run remotely and a frontend client that can run in multiple environments including web? I remember a long time ago pyCharm had the ability to run the interpreter remotely. What are the benefits of pulling more stuff in the backend? Is it in order to be able to run the editor in different environments?

Thanks for the quick answer! Are the examples generated based on the types? As in if a field is set as an integer, the tests will pass arbitrary integers? But still how does it know how to produce positive examples? Do you maybe have some documentation on this functionality?

Out of curiously, does HN use JS heavily? It feels quite lightweight, and it’s sort of a forum

I was looking in the site to understand the automatic generation of tests from OpenAPI but I couldn’t find it. Maybe it’s very obvious but I missed it. Does it mean that it produces test cases from the examples provided? Or how does it figure out the request response pairs?

I would definitely try to get something new before quitting. But that might be easier said than done if you’re already burning out.

I agree with most comments that it’s ok to prefer not being VC funded. Looking at what other possible outcomes you may have by taking an alternate route: if you take VC money and the company self-destructs, you might end up with considerable capital (depending on how things ended) at the expense of the business and your clients.

But that’s a gamble. If you think the current MRR is sustainable then I would personally stick with it (I assume you’re a solopreneur)

Would the same advice hold for demos of e2e solutions, say backend and frontend? At least in scrum, stories usually require input from many areas, and the demos usually center around the UI (since it’s the easiest to show).

In this case I assume at which point you talk about backend might be important. For example if it’s at the end, people will just drop off since “it’s boring”.

More and more I’m thinking of having open software that runs on client hardware and that sync via simple P2P connections. For example having some Evernote clone where the mobile app is in sync with the desktop app without using a centralized server.

My issue is being trapped in SaaS platforms with bad business models that threaten the existence of the service. Think of Instagram, Tiktok, Twitter, Facebook, heck even Uber and Uber eats clones.

Though I’m not into the crypto stuff. I’d be OK if banks kept centralized ledgers that these free-to-use privacy-oriented P2P services could use, for crowd sourced applications.

I really miss the feeling of the internet in the 90s and 00s but how could people work on these projects while still make a living?

How does something like react native fit in the mobile world? Are there cases where it could be used instead of straight swift or kotlin?

As others have mentioned, one can get some types in python through its own typing module and use a combination of tools for type enforcing and editor intellisense (mypy, pylance, etc.).

However what I’m missing is having more advanced type constructs. For example mapped types, lookup types, etc [1].

That said, it seems like python devs keep adding features over time. For example a couple of years ago it wasn’t possible to specify which fields of a dictionary are required and not required (it required putting required fields in one TypedDict and then subclassing it into another one that had the non required ones with total=False), nor it was possible to “spread” a TypedDict to declare the kwarg types in a function.

[1]: https://www.typescriptlang.org/docs/handbook/release-notes/t...

He’s a lecturer at MIT

I wonder how much classical education hinders this kind of people. I can see a point in networking, however creating a network of randomly assigned people doesn’t seem so effective. It might be better to create a network from clubs, etc.

I never created anything so impressive while young but I was always curious and building or taking apart stuff. I always felt school got in the way.

in the end I stayed all the way until PhD and later changed careers, but that’s a different story

Sometimes a bunch of us meet at the office to work together on something that needs a lot of back and forth, like a brainstorm session. And to socialize during lunch. If independent work is needed then we’re all at home.