Wow, you are not wrong. I just asked Gemini "how can I use ffmpeg to apply a lower third image to a video?" and it gave a very detailed explanation of using an overlay filter. Have not tested its answer yet but on its face it looks legit.
back
3 comments
It could very well be legit, but if you "have not tested its answer yet" the fact that it can generate something that looks plausible doesn't really tell you much. Generating plausible-sounding but incorrect answers is like the #1 most common failure mode for LLMs.
in recent usage, that only happens 10% of the time for me. Usually the results are grounded and so work usually fine.
Could you imagine if any other software failed silently and plausibly 10% of the time? It would never get off the ground. VC money is a hell of a drug
It's amazing --- I cut my teeth in software engineering with ffmpeg-related work 15 years ago, LLMs generating CLI commands with filters etc. is right up there with "bash scripts" as things LLMs turned from "theoratically possible, but no thanks unless you're paying me" into fun, easy, and regular.
Yesterday I asked it for a command to take a 14 minute video, play the first 10 seconds in realtime, and rest at 10x speed. The ffmpeg CLI syntax always seemed to be able to do anything if you could keep it all in you head, but I was still surprised to see that ffmpeg could do it all in one command.
I never found bash scripting disagreeable. I have thousands of scripts for both work and my everyday computer usage. I keep a ~/bin folder in my path where I place useful scripts that I have written for my own use. One thing that keeps me using bash for this purpose over Python or Ruby (which I use for serious programming) is that I can take a command line invocation that I manually constructed and tested and put it in a script without modification.
Me too! I am a data engineer so whenever I have pipeline jobs running, I have a script that monitors them. When the jobs finish, an audio plays stating the job has finished and its status. Thats just one convenient script out of dozens.
Makes life much more easier when I can play video games or read books without having to check status every 20 mins.
Though I haven't created as many as you have. Would you mind sharing some of them??
Thousands is probably hyperbole, but there are many, many!
One that I find regular use for when copy+paste does not work because I am either connected to a terminal emulation, vm, or something of the like is typeitforme. It takes the contents of a text file and sends them through the keyboard buffer after a few second delay (that allows me time to switch focus to the window I want the typing done).
I currently have it as a entry in my ~/.bash_aliasas file.
alias typeitforme='sleep 3 && xdotool type --file '
This works in stock Ubuntu Linux. You can check out the xdotool documentation for ideas how to refine it to do more.
"Have not tested its answer yet but on its face it looks legit."
That's LLMs for you
Fair point, but I only had limited HN commenting time budgeted, not getting a video set up to test this idea. I did confirm that the overlay feature exists via the official ffmpeg documentation.