gptel-org-tools update.

Edit: there's some kind of issue with @Codeberg@social.anoxinon.de which prevents the link from working (returns 404). The old (but up to date) repo is here: git.bajsicki.com/phil/gptel-or…
1. Cloned to codeberg.org/bajsicki/gptel-or…, and all future work will be happening on Codeberg.
2. Added gptel-org-tools-result-limit and a helper function for it. This sets a hard limit on the number of characters a tool can return. If it's over that, the LLM is prompted to be more specific in its query. Not applied to all tools, just the ones that are likely to blow up the context window.
3. Added docstrings for the functions called by the tools, so LLMs can look up their definitions.
4. Improved the precision of some tool descriptions so instructions are easier to follow.
5. Some minor improvements w/r/t function names and calls, logic, etc. Basic QA.

Now, as a user:
1. I'm finding it increasingly frustrating that Gemma 3 refuses to follow instructions. So here's a PSA: Gemma 3 doesn't respect the system prompt. It treats it just the same as any other user input.
2. Mistral 24B is a mixed bag. I'm not sure if it's my settings or something else, but it fairly consistently ends up looping; it'll call the same tool over and over again with the exact same arguments. This happens with other models as well, but not nearly as frequently.
3. Qwen 2.5 14B: pretty dang good, I'd say. The Cogito fine-tune is also surprisingly usable.
4. Prompting: I have found that a good, detailed system prompt tends to /somewhat/ improve results, especially if it contains clear directions on where to look for things related to specific topics. I'm still in the middle of writing one that's accurate to my Emacs set-up, but when I do finish it, it'll be in the repository as an example.
5. One issue that I still struggle with is that the LLMs don't take any time to process the user request. Often they'll find some relevant information in one file, and then decide that's enough and just refuse to look any further. Often devolving into traversing directories /as if/ they're looking for something... and they get stuck doing that without end.

It all boils down to the fact that LLMs aren't intelligent, so while I have a reasonable foundation for the data collection, the major focus is on creating guardrails, processes and inescapable sequences. These will (ideally) railroad LLMs into doing actual research and processing before they deliver a summary/ report based on the org-mode notes I have.

Tags:
#Emacs #gptel #codeberg #forgejo #orgmode #orgql #llm #ai #informationmanagement #gptelorgtools

Now that #GoToSocial v0.19.0 is out, we're going to start the process of transferring our coding workflow from Github over to #Codeberg (@Codeberg), to get away from Microsoft's endless parade of crapola.

This means that some time in the next week or so, we will "lock" the Github repository so that only superseriousbusiness.org members can create new PRs and issues etc (similar to what we do when we're on holiday).

Then, we will migrate the repository over to Codeberg, using Codeberg's migration functionality to keep issues, pull requests, comments, etc intact.

Once we've validated that everything is in order, we will mark the Github repository as a read-only mirror of the Codeberg repo, and start using Codeberg for issues, pull requests, CI/CD etc.

To avoid breaking people's automated builds that rely on that Github repo, we are planning to keep it as a read-only mirror for a while (eg., 6 months or so, not sure yet) . At some point though we will fully remove it, to avoid contributing our code to Microsoft's "ai" garbage. But we will warn in advance about this so that people can update their scripts and stuff in time.

There will probably be some turbulence in our CI/CD pipelines and workflow and whatnot during this transition period, but nothing that should affect end users of the software: releases up to v0.19.0 will remain available on Github and Docker as normal, and snapshots should remain up to date on our Minio S3 bucket and Docker.

We're looking forward to the move, and to putting the sloth in its new home and seeing it wander around sniffing the furniture and exploring Thanks for reading!

new release of fj.el, an #emacs #forgejo / #codeberg client. it adds:

- pagination of most views
- generic next/prev page functions
- generic reload engine, making it trivial to reload a specific view while changing a single parameter
- basic milestones support (add to existing issue, add while composing, display when viewing)
- new macro to handle optional parameters (in fedi.el), makes handling the full API much easier
- render item markdown using instance API (removes need for markdown binary)
- use server default for limit arg always
- mark notification read upon viewing item
- toggle notifications view read/unread/all
- cycle notification type
- list issues by milestone/label
- render PR review diffs and discussion in PR view
- display/add/remove reactions
- implement font-locking of issue numbers/user handles while composing
- add labels while composing
- fix multibyte encoding errors
- handle repos that you do not own yourself (via git)
- quote region in reply buffer
- cycle list own items by state and type (accross all repos)
- handle auth-source for encrypted access tokens
- support repos on different Forgejo instances via .dir-locals.el settings.

there's much need for improvement still, but it's coming along well.

Oh and who’s behind it? The fucking fascists, of course. Because, of course. Fucking fascists.

“Codeberg is currently suffering from hate campaigns due to far-right forces … In the past days, several projects advocating tolerance and equal rights on Codeberg have been subject to hate attacks, such as massive spam of abusive messages in their issue trackers.
…

We will not be discouraged in our fight against far-right ideologies. They are currently on the rise in many parts of the world, and we believe it is important to protect all kinds of marginalized groups. However, if you believe this does not affect your project, you are wrong. Far-right forces pose a threat to all of us.

Extreme right forces actively target members of our communities and discriminate based on ethnicity and gender, political background, sexual orientation, disabilities, nationality and faith. However diversity is an important asset in free/libre software communities and it is what makes our software great and development productive.

By targetting some of our most active translators, nicest designers, best developers and all other motivated contributors, they are hurting the free/libre software ecosystem as a whole.

Don't be fooled if right-wing forces promise to "promote open source" in their political agenda. This has nothing to do with the values of our movement! This is about national patriotism and protectionism, and they will happily accept splitting our community on their way.

We all know that the free/libre software ecosystem won't work this way. Every human is an integral and equally important part of it, and targetting some of our community members threatens the ecosystem as a whole.

Let's together stay strong and united against the emerging threats. We stay strong against discrimination of all kind, including but not limited to sexism, transphobia, homophobia, racism, antisemitism and ableism. And we hope that you all join us for this mission.

A big shout-out to all the projects that collect facts and resources against hate and discrimination and that have been the primary goal of these attacks. Support them if you can.

If you ever considered supporting the fight against right-wing forces, for example by joining political movements and parties, organizing protests or getting involved in online communities with this goal - now might be the best time to move ahead.

Again, we apologize for the disruption of your work and the abusive content delivered to your mailboxes. We are doing our best to contain the situation.

We'd like to thank everyone who signaled their support in the current situation, via e-mail, Mastodon and in Matrix chats. This means a lot to us.

Thank you for your trust and support!
Your Codeberg Public Relations team”

blog.codeberg.org/we-stay-stro…

#Codeberg #CodebergUnderAttack #fascism

So you want to fight fascism and make a difference right now? Do this:

Join Codeberg e.V. – become part of the organisation; part of the resistance – and actively support this valuable and ethical EU-based anti-fascist¹ git hosting alternative that’s under attack by fascists as we speak:

join.codeberg.org

Seriously, do it now.²

Volunteer and/or donate if you can.

Let’s send Codeberg’s membership through the roof and show these fucking fascists that when they push we push back twice as hard.

When you do, please add your voice to this thread so others can see.

I rarely ask this but please repost this far and wide.

¹ blog.codeberg.org/we-stay-stro…

² If you’re having trouble reaching the server, it might also be under attack. You can check the current status of all Codeberg’s servers from status.codeberg.org/status/cod…

#Codeberg #CodebergUnderAttack #SupportCodeberg #antifascist #EU #european #FOSS #freeSoftware #git #hosting #fediverse #mastodon #community #callToAction