How do you measure positive and negative societal impact of this technology?
I find mobile phone face unlock so useful, giving every citizen the power to use face recognition could be better than a few people, robots that identify someone and give them lifesaving medication are great (but the opposite, robot assassin can also be created). I guess it comes down to good people building good tools. Humans are generally kind and empathetic
It'll be a net-positive for things like that, passports might be rendered useless in the future since they're already using facial recognition at Customs. It can help solve crimes given it can recognize faces using CCTV footage amongst other things.
There is a certain level of distrust since it can be abused and people think it will lead to a dystopian police state.
It depends on the tool you use and I guess the use case too. Some are language model agnostic like aider in the command line, I use sonnit sometimes and then 4o other times. I wonder if or when language models will become highly differentiable. Right now I see them more like a commodity that are relatively interchangeable but that is shifting slightly with other features as they battle to become platforms
I built an email notifier for alerting me when there is a popular news article which I defined by X number of comments. I think forums like hacker news are a really cool indicator for an emerging technology, an opportunity, or some exciting tech story. I get a tonne of false positives though, mostly articles that are inflammatory get sent to me. However, crowd strike triggered it as did the release of chat gpt so there are some hits
Technology like this will create huge selection pressures against desktop apps if it becomes easy for people to reverse engineer, remove payment mechanisms, and then freely distribute. Wouldn’t you think?
Delinking isn't a miracle technology in that regard, you still need to put in the work to reverse-engineer the artifact.
It does allow a couple of nifty tricks, like pervasive binary patching (if the program is chunked into relocatable object files, then you're no longer constrained by the original program memory layout when patching/replacing stuff). It's also useful for decompilation projects, where you can reimplement a program one piece at a time until you no longer have binary pieces left and still create a fully working program at each step (you don't even need perfectly matching decompilation since the linker will mend stuff back together anyway).
Freely redistribute, I don't know. If a program can extract code from another one, a program can detect that code. It looks similar to virus signatures. A company with some IP would run the detector on the software of competitors.
Did you get ChatGPT to help you write this and is one of your main use cases generating bash commands? (Code writing code, there’s a total feedback loop happening here) I did too but wrote mine in Python. A compiled language would be much faster and using their streaming function I’m yet to implement. Any other ways you can increase the speed?
I don't think language would matter much as the primary delay is getting a response from OpenAI apis. The advantage of Rust (or other similar language) is you get a simple single executable. Otherwise Python is completely fine.
I basically use this any time that I am in the terminal already and have a question that I think could reasonably be formatted fine there. I'd say, I use it for 30% (conservatively) of my ChatGpt queries. Naturally the web interface is way better for lots of stuff, but sometimes breaking out of a terminal session is jarring.
- Oracle Cloud Free Tier[1] for a Ubuntu VPS (4 ARM cores, 24 GB RAM). Surprisingly pleasant and reliable, given who's offering and for how much ($0). It used to be on DigitalOcean, until they kept screwing up their FreeBSD support and bricked my machine twice.
- Caddy[2] web server with Let's Encrypt certificates, working as reverse proxy.
The rest is a very lazy 2010's solution:
- A Go server for HTTP (static files, uploads, maintaining server-sent-event channels). It also reads and writes events in a custom format to a local socket, for the interactive parts.
- A Python server for the widgets, communicating with the Go server.
- Source code edited manually in-place (SSH or SSHFS[3], with Git) and restarted as needed. I know, I know, awful practice. But as I'm the only user, uptime during development is not a concern.
- Startup is handled by a @reboot cronjob and a bash one liner.
- Text files for "structured storage" (RSS feed items, authenticated sessions, mapping of uploaded file names).
As horrible as it might all sound, it has survived ten years and two cloud vendors. Nowadays I might containerize it, or rewrite as one Rust server, but I think I made the right choices at the time.
