I would love to see the push10k on android playstore/fdroid.
It looks so inviting and motivating that I searched for equivalent alternatives for android but found none!
Would you, maybe, publish it there as well, pleeeease?
Could we please stop putting price tags on 15-commit repos?
It's just crazy that every idea, created with ai, now costs 10$ or more per month, despite it costs 5$ to create.
Yes. In my case (and I guess everyones usecase is subjective) my system prompt states to read the AGENT.md file when possible.
On a new project I usually set up the context of the model (language to use, reason of the product/prototype, etc.) and then I tell the LLM to write a AGENT.md, STATE.md and ROADMAP.md.
I don't tell the LLM what's in there because the model has it's own directive and flavor what should be in these files. The models already know the purpose of these files by themself! On a new session, I let the agent read the markdown files in order to continue with the work. Before a session ends, I let the LLM update the markdown files.
Maybe one word of caution: don't switch models - it's like putting another person on a working station and ask them to continue the work of others.
I do something similar. I don’t necessarily call it AGENT or STATE and every project has its own files. I have architecture documents that accompany change log descriptions that load technical knowledge that the agent can readily use.
I find it also necessary to have a principles document outlining the particular problems that the software is supposed to solve and guard rails to not cross. I call it promise driven development.
Yea absolutely, but man, where to even start, it is very specific.
Fundementally I didn't use any wrappers like unsloth or axolotl, although I have used the latter before a year or two back and it was good, but I needed something very very custom. I also wanted the whole fine tuning pipeline to exported OpenVino model to be seamless.
I heavily leaned on codex, claude and some manual sleuthing around the internet to understand what I needed. I'd played about with QLoRA finetuning with axolotl before and felt most comfortable with that. So I needed to keep everything as stripped down as possible and figured I can just utilise the 3 main huggingface libraries (transformers, peft and datasets) and also bitsandbytes (as suggested by claude to quantize the model to keep this working on my GPU) along with some custom scripts generated by claude/codex (each cross referencing each other) that will do the different stages of the training run.
The next part was the data. Obviously didn't have access to thousands of meetings and associated output documents but I did have a 3090ti sitting there and a codex subscription. So I set about working out what format I needed the data in (many thanks again, to claude/codex) and started generating hundreds of different transcripts, different amounts of speakers, content, tones, subjects, spelling mistakes - like all the different things you could think a meeting would have. Then it's a case of actually generating a good meeting document off the back of the transcripts and creating the "gold standard" that we'd use.
I'm going to gloss over a lot here as I'd rather not detail it as it relates to some propriatary stuff that I had to work through, but you basically pair the transcripts together and run the training.
At the verification stage, there was pretty much 3 things:
1. "just" do some regex string matching to see if there's any of the source transcript key facts in the output to ensure fact preservation. Same with owner fabrication (who said what), I don't want something attributed to someone when it wasn't them that said it and then finally markdown validation.
2. Using codex/claude to validate the transcript and output from the model - I used the latest frontier models, probably overkill for my task, but they were good at the job
3. Finally me going through some actual recordings of myself, groups, meetings and manually verifiying the output
So a fair bit of work, and for context I'm on version 10 now, so it's been a journey!
Don't worry, they'll make it borderline free at the start, let everyone integrate into their org, and then jack up the price after just like they do with everything else.
And when you had a tool call that asked the user for the next step, you could easily run a whole day with 4c. Guess how the people did 5k $ worth of token with 100$ spent.
Microsoft just introduced another virtual currency to tinker with.
The comparison with adult webcam services are actually quite fair: you buy the currency, spend it and watch the actor doing it's job... but now in the tech ecosystem.
The readme tells me much, but doesn't say anything helpful.
I would love to have the authors of the ai tools would at least put a bit love into the explanation...
Anyway, I learned about https://handy.computer which is nice!
Combined with Whisper+ from F-Droid and you have a nice way to input voice as text from everywhere.
I thank you to take my feedback seriously and I read it again.
It's now much clearer what the project is about!
May I clarify my previous feedback a bit more: I was trying to wish for handcrafted readme. Vibecode the product is fully ok, but people buy ideas from humans. Once you've let build your idea, take your time and write the humans, that should use it, a handwritten letter.
For what other reason than another himan should use it, did you uploaded it to a public repo and post about it?
Oh I see. Perfect. I was just fine tuning and looking for feedback and I I have now tested more in my own business and but I haven't made a big splash, I just posted here and that's it. It was just like to connect some initial feedback.
Yeah about writing it manually that would be good I was just trying to make it better since I'm not a native English speaker so was just trying to make it a bit better. :)
reply