Anyone else have there own git and have a bunch of programs / mods / addons that we dont share? I would release what ever I have fully open source but then I worry if someone uses it for something stupid or it fucks up someones shit. Is someone now going to come after me? Would anyone even want it? Maybe someone could learn something off it ?
Is it even worth releasing stuff that ai has made me in ways that I want it ?
Linus and Luke have an opinion on that:
i have a bunch of crappy non-vibe coded bash scripts i wrote for different things. most are short, i prefer to use python longer scripts, its generally faster than bash.
stuff like :
#!/bin/bash
for i in $(ls /var/lib/flatpak/exports/bin/); do a=$(echo $i | awk -F "." '{print $NF}');ln -s /var/lib/flatpak/exports/bin/$i /home/user/bin/$a 2>/dev/null;donethis one links flatpak programs to the ~/bin folder (which i added to my $PATH) . This way i can launch a flatpak program from the cli without typing
flatpak run program.name.blah.blan.blah.something.org, just the program name.
I’m honestly trying to force myself to code simple things again. There are some home lab deployments I want to do in Ansible. I could drop my prior ansible playbooks in a project and write a spec for how I want them adjusted. It would take the AI demon maybe 5 ~ 10 min and it might take me an hour or more.
I’m trying to interview for jobs right now. I want to code myself for that, yes … but also .. just for my own cognitive abilities. I don’t think this path for over-reliance on LLMs is sustainable, but I don’t know where we go from here either.
I do release all my my AI spec coded stuff though, all using the FOSS project that’s very against pure AI coding too:
Most open source software comes with the usual disclaimer in its license to best-effort waive any responsibility.
THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Why / how would they? As long as your AI did not reverse engineer any proprietary software or includes a software to pirate stuff from Disney and co. you are most likely fine. Put a MIT license in each repo and you are good to go.
In my experience probably only a small number of people, declining with the number of AI generated projects. But I think knowledge should be free and in almost all cases it does not hurt to go FOSS.
Edit: btw. @starshipeleven quote is from the MIT license
I don’t plan on publishing anything. It’s mostly stuff geared directly towards my own use cases anyway. Stuff like scripts to deploy things into containers etc. Nothing someone else can’t whip up themselves.
Only thing that is even accessible at all is this: https://tls.obsidian.place/ - fingerprint analysis tool. And that’s specifically because it’s low risk on the “fucks up someone’s shit” front as you say.
But what if the (short sighted) employer does not want you to? Because this:
It is all they see.
It sure isn’t.
Most likely more LLM as long as “line goes up”
GPL has similar statements also in FULL CAPS
16. Limitation of Liability.
IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
and here from the 3-clause BSD license
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
So OP is free to choose between the major copyleft or permissive licenses
So I’ve vibe coded a few little utilities that solve specific little problems for me. Things like automating my timesheet capture from actual activity on my work laptop. Some finance tools (luckily my bank has an API customers can use). I’ve also built a tool for a local non-profit that’s working really well and they really don’t care that it’s vibe coded - took little of my time and its free to them.
There’s a bigger thing coming though, and this to me is a signal:
We need a term for this, for now I’m calling it vibecloning. On the one hand as a dev, it signals that products have no practical barrier to prevent this. If it can be done with Photoshop, that’s quite a high bar and then I’m not sure what kind of software would be beyond reach for someone motivated enough to spend the tokens.
Since Jan. when I began the LLM assited software generation we’ve made:
flexd - webserver
pkb.rs - personal knowledge base
cargo-vuln-check - deps scanner
frontpage - my homepage is offline
OSC - the spec that drives the software
vsent - vuln sentinal
sandbox-ctrl - A packaged version of my workstation isolation scheme
verify-down - A download verification automation, any SHA/SIG/ASC file is pulled and verified with a ledger.
Ongoing work on two massive projects alongside of those above.
It’s a bold strategy Cotton, to openly admit you reverse engineered a piece of Adobe software as their EULA states:
“You will not reverse-engineer, decompile, disassemble, or otherwise attempt to discover the source code, data representations or underlying algorithms, processes, methods, or any other portion of the Software.”
I’m guessing whoever admitted to it, doesn’t really like their house or savings, as it’s pretty much a slam dunk if Adobe sues them.
That is if you agree to the EULA - in what terms you would “not agree” and be able to do this? There’s going to be a lawsuit for sure, and if there are ways for ArtCraft to win, I just hope that there is enough money to survive the trial.
I would say use AGPLv3. I’m still a hard-core GPL advocate plus I like the idea of poisoning the entire pool of generated code.
I keep all my LLM generated code on my forgejo instance:
I’ve got the machines to make a dual-stack IPv4/IPv6 status board:
Also made an image sorter, at better web interface for a typical PS4 jailbreak, a tool for archiving Teams messages, a grammar checker for Pulsar (formerly Atom) editor (uses a languagetool container), and a better interface to a ComfyUI server for the console.
Prove I accepted the EULA 
Shit we listing our AI creations now? I made these available on my github with the same username:
+1 for putting the bespoke code into a private forgejo instance
Licensing generally will shield anything used by others aside from gross negligence. The same was true for the hand rolled, artisanal code projects of yore. Nice to make things public, but if it’s super specific, like I had an LLM write a script to migrate my setup from one PC to another (it was a one time use thing), probably okay to keep it for yourself.
My guess is that after awhile, most private projects with value in reuse, will get refined or even scoped down. Once you get it to something that’s more generalizable, it’s probably reasonable to share. Ofc you could also carve out small chunks of a larger project that might be useful too. That could help reduce the potential blast radius if you’re concerned about what damage it could do