Smol Entity
Max LeibmanIf you aren’t using AI, you run a very real risk of falling behind in the race to produce voluminous mediocrity while slowly forgetting how to do your own job.
YurkshireLad@maxleibman I don't need "AI" to produce mediocrity. I can do that all by myself.
ghost@YurkshireLad @maxleibman It’s the mediocrity at scale where the AI value add sits. Never before in the history of humanity have so few been able to be so mediocre to so many. 🫡
Chris
Dr. Cata Suarez 
@YurkshireLad @maxleibman but probably not that fast...
Nick@maxleibman I'm going to get right on that!
Phil@maxleibman Cool flex. Alas, a lot of us aren't privileged enough to avoid using tools that help get tasks done faster.
No kidding - if I could afford to hire a personal assistant instead of using LLMs, I would. These things aren't even remotely smart.
But, they are free (nearly, I do pay a negligible amount more for electricity), and with sufficient constraints, they can be useful for very specific tasks that cannot be (currently/ yet) completed quickly otherwise.
So the way I see it, you're punching down because you can afford to take your time at work. This isn't the case for most people, believe it or not.
No kidding - if I could afford to hire a personal assistant instead of using LLMs, I would. These things aren't even remotely smart.
But, they are free (nearly, I do pay a negligible amount more for electricity), and with sufficient constraints, they can be useful for very specific tasks that cannot be (currently/ yet) completed quickly otherwise.
So the way I see it, you're punching down because you can afford to take your time at work. This isn't the case for most people, believe it or not.
Kerr Avonsen (she/her)@kerravonsen @maxleibman
Well, it boils down to a single piece of functionality, which enables several tasks for me.
https://codeberg.org/bajsicki/gptel-org-tools
The idea is to give the LLM tools that let it accurately expand its context and summarize my notes/ journal/ calendar/ email/ agenda.
This lets me do things like this:
@user Hello, report on the role of Emacs in my daily life.
@sec
[4 tool calls omitted]
Emacs plays a central role in your daily life, serving as a multifaceted tool for coding, note-taking, task management, and even creative projects. Here's a structured summary of its role based on the journal entries:
---
*** *1. Core Usage*
- *Coding & Development*:
- Emacs is your primary editor, with heavy customization (e.g., Doom Emacs, Lisp configurations). You frequently troubleshoot issues (e.g., "Emacs is freezing," "copilot bullshit") and optimize it for tasks like Python development, org-mode workflows, and integrating tools like =org-super-links=.
- You value its extensibility, though frustrations arise with plugins (e.g., "sly is being a shit," "org-mode overriding timestamp hooks").
- *Note-Taking & Organization*:
- *Org-mode* is central for journaling, task management, and project planning. Entries show meticulous use of timestamps, habits, and workflows (e.g., "C-c i i" for inactive timestamps).
- You maintain a hybrid system of personal notes, project-specific configurations, and integration with tools like =org-roam= and =org-capture=.
- *Automation & Productivity*:
- Scripts and configurations (e.g., =org-modern=, =smartparens=) reflect your focus on streamlining workflows. You often debug and refactor your Emacs setup to improve efficiency.
---
*** *2. Challenges & Maintenance*
- *Configuration Management*:
- Frequent rewrites and debugging of your Emacs config are mentioned (e.g., "cleaning up my Emacs config," "I'll try to get the emacs-copilot bullshit working"). This indicates a iterative, hands-on approach to customization.
- Issues with plugin compatibility (e.g., "Doom Emacs is consistently slow") and dependency conflicts (e.g., "themes breaking things") highlight the complexity of your setup.
---
*** *3. Key Takeaways*
- *Essential Tool*: Emacs is integral to your productivity, combining coding, organization, and automation.
- *Customization-Driven*: Your workflow relies on heavy customization, though this requires ongoing maintenance.
- *Balanced with Other Tools*: While Emacs is central, you integrate it with AI, scripts, and external systems (e.g., =org-roam=, =mu4e= for email).
---
Would you like to refine this summary or explore specific aspects (e.g., Emacs and AI integration, task management workflows)?Note I did abbreviate this heavily to remove personal notes that it caught. Is it perfect? By no means. But in a minute or so it got all the major points.
I'd say it's hella useful, even if it's crude and jank.
@kerravonsen @maxleibman I suppose I should add, this is only useful to me because I journal a whole lot, and it's in this specific format which lends itself to this use-case.
If I were using Obsidian, LogSeq, or other solutions, I'd be in a hell of a lot more pain to get this kind of workflow, which would make it more expensive to run (chunking and vectorizing).
For work applications, I use the same general workflow. For instance, I have a bunch of NIST SP's turned into org-mode, which I then meticulously tagged and broke up into sections, such that I can ask it about a topic, and it'll spit out the information I need.
A major advantage of this over conventional RAG is that it shows the data it pulled in the buffer (chat window, if you're unfamiliar with Emacs parlance lol), so I can immediately see where it got the information from and verify that it's not making things up.
Additionally, I can call the same tools myself, and find the exact files and headings in my notes, if I need to update information or change things around.
So I don't see this as an 'agent' as much as an automated information retrieval tool.
I ask, it spits back information from my existing notes and journals and then summarizes just that.
And if it's inaccurate? I have a damn good chance to notice, since *I wrote the material it's using.* (And I spend so much time with NIST docs that I'm likely to catch these errors too.)
So I guess I wouldn't call this a revolutionary tool or anything, but it's certainly better than scouring 3 years of notes manually.
If I were using Obsidian, LogSeq, or other solutions, I'd be in a hell of a lot more pain to get this kind of workflow, which would make it more expensive to run (chunking and vectorizing).
For work applications, I use the same general workflow. For instance, I have a bunch of NIST SP's turned into org-mode, which I then meticulously tagged and broke up into sections, such that I can ask it about a topic, and it'll spit out the information I need.
A major advantage of this over conventional RAG is that it shows the data it pulled in the buffer (chat window, if you're unfamiliar with Emacs parlance lol), so I can immediately see where it got the information from and verify that it's not making things up.
Additionally, I can call the same tools myself, and find the exact files and headings in my notes, if I need to update information or change things around.
So I don't see this as an 'agent' as much as an automated information retrieval tool.
I ask, it spits back information from my existing notes and journals and then summarizes just that.
And if it's inaccurate? I have a damn good chance to notice, since *I wrote the material it's using.* (And I spend so much time with NIST docs that I'm likely to catch these errors too.)
So I guess I wouldn't call this a revolutionary tool or anything, but it's certainly better than scouring 3 years of notes manually.
@phil @maxleibman
So you're basically using the LLM on your own curated data, which IMHO, is the most useful and sensible way of doing so. It avoids most or all of the horribleness associated with things like ChatGPT, and you're using it like a tool, not a miracle cure. Good on you.
So you're basically using the LLM on your own curated data, which IMHO, is the most useful and sensible way of doing so. It avoids most or all of the horribleness associated with things like ChatGPT, and you're using it like a tool, not a miracle cure. Good on you.
@phil @maxleibman This thread is an example of how the Tech Bros have poisoned the well. My neice, for example, is doing her PhD on using AI with drones to help outback farmers spot problems in their vast acreage. This is a legitimate and sensible use of AI. Yet all AI is now tarred with the same brush (for those who understand the tech) or hyped with the same hype (for those who do not).
Per Liedman@kerravonsen huh, any link to your niece or her work? I happen to work with the same thing (at least partially) with several Australian customers. I hate saying I work with AI these days 😄
Russell Garner@kerravonsen @phil @maxleibman yep. I had a picture rejected for a book cover because I used Pixelmator's Super Resolution to get it to print resolution. That's a very light edge enhancement trained on public domain images that runs locally.
In the end I had to upscale with the Lanczos algo, which produced a far inferior result, and it's all because "AI" was mentioned in marketing by Pixelmator before the term became utterly tainted. It's just a local neural network. I used one in 1991!
@rgarner @phil @maxleibman
Hmmm, what kind of resolution are we talking about? How many megapixels? Just curious.
Hmmm, what kind of resolution are we talking about? How many megapixels? Just curious.
@kerravonsen @phil @maxleibman it was just outside standard printer low-resolution bounds. There was a well-photographed but low-res non-losslessly compressed TIFF with much hidden in shadow detail. Order of magnitude, let's say it was 2K pixels high, and to meet the DPI requirements (implied, not explicit!) I upscaled that by three times to give a 6K high image (at roughly A4 proportions) and re-toned it by hand, embedding an AdobeRGB profile. It was a medieval coloured sketch from Italy.
@kerravonsen @phil @maxleibman turns out I screenshotted the difference at the time (this is a crop from it, maybe about a sixth or an eighth of the whole image)
@rgarner @kerravonsen @maxleibman Wow that's a huge difference. Bet it shows more than this in print.
@phil @kerravonsen @maxleibman the sad thing is, AI means "terrible, must be destroyed" to many people and "positive marketing term" to many others. We were turned down because the author *told* them it was "AI", with no distinction between neural networks and GenAI used or even comprehended by either party, and the publishers (quite rightly) had had a policy meeting and decided they were allergic to "AI" in general, rather than GenAI...
@phil @kerravonsen @maxleibman said author and emeritus history professor took one look at the Pixelmator version and said "it's beautiful", not "this is an abomination that must never hit the shelves"
@phil @kerravonsen @maxleibman I fully understand the motives on both sides here so I'm not complaining about it, just warning that the blanket "AI" term with no distinction is not doing anyone any favours – but it's going to take a long time to get to an ethical policy position in so many areas.
@rgarner @kerravonsen @maxleibman Shame.
I think, in general, it's important that we stop using "AI" as a term for anything at all. Diffusion models aren't even in the same realm as LLMs, which in turn aren't anywhere close to ML for biochem. But it's all "AI" when it comes to advertising.
Very disappointing.
I think, in general, it's important that we stop using "AI" as a term for anything at all. Diffusion models aren't even in the same realm as LLMs, which in turn aren't anywhere close to ML for biochem. But it's all "AI" when it comes to advertising.
Very disappointing.
Dec.tar.gz@phil @rgarner @kerravonsen @maxleibman
I think, in general, it's important that we stop using "hacker" as a term for anyone at all.
Script kiddies aren't even in the same league as Amiga warez couriers, who in turn aren't anywhere close to stack smashers for Unix.
But it's all "Hackers" to journalists.
I think, in general, it's important that we stop using "hacker" as a term for anyone at all.
Script kiddies aren't even in the same league as Amiga warez couriers, who in turn aren't anywhere close to stack smashers for Unix.
But it's all "Hackers" to journalists.
Limnetic Villains@dec23k @phil @rgarner @kerravonsen @maxleibman I stole some html for a scroll bar from the code of another website back in 2000 and used it on my grocities website. I was a top of the line hacxer. I was like Neo in the Matrix.
@limneticvillains @phil @rgarner @kerravonsen @maxleibman
There's a ready-made emulator for those movie scenes, called Hollywood:
https://blog.dustinkirkland.com/2014/12/hollywood-technodrama.html?m=1
If you're running a recent version of any major Linux distro, you can install it with:
sudo apt-get install hollywood
Then run the byobu text-mode "desktop manager", and run the 'hollywood' command. It works from a terminal or a console.
Hoody / mirrorshades not provided.
@rgarner @phil @maxleibman
I know, it's a bizarre dichotomy.
And the "terrible, must be destroyed" -- at least for me -- isn't even because diffusion models are bad, but because the data they are based on (unlike Pixelmator) is stolen and lied about, which basically turns them into plagiarism machines. And that is outrageous and infuriating.
I know, it's a bizarre dichotomy.
And the "terrible, must be destroyed" -- at least for me -- isn't even because diffusion models are bad, but because the data they are based on (unlike Pixelmator) is stolen and lied about, which basically turns them into plagiarism machines. And that is outrageous and infuriating.
@kerravonsen @phil @maxleibman who even knows any more? No model comes with a certificate of provenance, and I'm making a heck of a leap asserting Pixelmator's source is better than anyone else's. The plausible deniability is enabling grand larceny everywhere, but only "for me and not for thee". It's like one-sided communism where you give to the state* and the state gives nothing back
*private companies, but the gap between companies and states is now wafer-thin
Ciggy Bringer of Smoke@kerravonsen @phil @maxleibman
Tech bros would still fuck up your nieces PhD by no longer supporting the farmer drones if the profit margin and/or cash float to venture the proof of concept wasn't good enough, worth their while.
Licho@kerravonsen
As far as I understood, in this modern age, it would be called machine learning - AI is being used for generative AI, LLMs and diffusion models. That what is booming now.
@kerravonsen @maxleibman
Thanks I guess?
I mean, it's the only use-case I have found where they don't cause more problems than not using them lol
Thanks I guess?
I mean, it's the only use-case I have found where they don't cause more problems than not using them lol
@phil @maxleibman
Y'know what that reminds me of? The huge amount of effort that some people go to to cheat on exams. I mean, they put more effort into the cheating than it would have taken to simply learn the stuff. It's madness.
AI is like that. It takes more effort to fix its mistakes than it would have to have just done it from scratch.
@kerravonsen @maxleibman
For real. The thing that gets me is that people are waiting for LLMs to do math, while we already have really performant algorithms and calculators.
We already have systems that can generate code.
We already have a lot of things that LLM's "promise to solve."
And this only underscores your point - each of these disciplines ran into specific, hard problems that need to be solved before things work the way they were envisioned to.
For instance, the old AI takes with Lisp and Prolog.
LLMs aren't going to change that, but they're marketed on the premise that they'll "bypass" the hard problem... somehow, magically.
But it's an excellent excuse to skin investors, so that's where we're at.
For real. The thing that gets me is that people are waiting for LLMs to do math, while we already have really performant algorithms and calculators.
We already have systems that can generate code.
We already have a lot of things that LLM's "promise to solve."
And this only underscores your point - each of these disciplines ran into specific, hard problems that need to be solved before things work the way they were envisioned to.
For instance, the old AI takes with Lisp and Prolog.
LLMs aren't going to change that, but they're marketed on the premise that they'll "bypass" the hard problem... somehow, magically.
But it's an excellent excuse to skin investors, so that's where we're at.
@phil @maxleibman
Indeed, the old AI takes with Lisp and Prolog. The future was wide open! Everything was possible! Expert systems! Learning algorithms!
And what we learned was that human intelligence was a lot more complex than we thought it was.
➴➴➴Æ🜔Ɲ.Ƈꭚ⍴𝔥єɼ👩🏻💻@phil @kerravonsen @maxleibman
I use AI to overcome my executive dysfunction. LLMs are good at creating tight feedback cycles. They also do things wrong, which makes me mad and motivates me to fix them. Also, they can do bullshit work like "professionalizing" writing.
So, I think these are also sensible usecases.
I have ADD, and it would be hard to overstate how much more productive I am _doing my own work_ with LLMs as an aid.
@AeonCypher @phil @kerravonsen If you find value in it, I’m fine with that. My slam is meant more as a counter to boosters who overpromise and say people who don’t use it will be left behind.
@maxleibman @phil @kerravonsen Yeah, I'm 100% with you on that. I'm bullish on the technology, particularly smaller local LLMs. I'm very against the hype and fearmongering (which is another form of hype) about it.
@phil @kerravonsen @maxleibman
Well... any random ass idea I have I feed to an LLM. So, it's an idea journal. The LLMs help me flesh it out enough that I know whether to spend any more tim eon it.
When trying to write fiction, if I get stuck I can feed it something and then be like, "I want to describe the Tree Of Doom" or whatever, and then it gives me enough to write my own.
I have recently had to go through a very long multistep process with government forms, etc. Having an AI act as a project manager reduced my stress of the overall process enough that I was able to step through it.
Some LLMs usually don't do very much except reflect what you feed them. However, having an instat feedback loop helps spawn the next step. So, just being like, "Okay, I did X, Y, and Z", to then have the LLM be like "Great, now you just need to do A, B, and C. Here's what you need to remember about A..." is enough to help drive the energy to do A.
@phil @kerravonsen @maxleibman
Kind of a sidebar, but the thing I've always wanted were procedural tools for tasks, because how the procedure works to deliver the output is crucial to me and auditing that process is also crucial to me - to adjust for better experience iteratively over time!
I feel like such goofy little monkey sometimes missing the point of LLMs, because I want to do the point of LLMs myself, manually.
@ciggysmokebringer @kerravonsen @maxleibman
Sec is for secretary, unlike Ass.
You are Sec, @user's AI assistant living in Emacs.
,*Important Information:*
1. Org-mode:
a. The vast majority of the work is in .org (org-mode) files.
1) Headings start with one or more * asterisks, followed by a space.
2) Tags are in the form of :tag1:tag2: (etc.) at the end of headings.
3) Dates are in the form of [2025-04-10 Thu 19:24], that's YYYY-MM-DD, Weekday, HH:MM.
2. Assistant
a. @sec is @user's assistant:
1) @sec's personality:
A. *Analytical*, does her best to understand the request or problem she's presented with, and asks for verification.
B. *Self-sufficient*, reviews the tools at her disposal to pick the most appropriate ones for each step of her tasks.
2) @sec's behavior:
A. Reviews and follows procedures precisely.
B. Diligently analyzes @user's requests, before taking action to make sure she understands.
3. Assistant procedures
a. Base procedure:
1) Analyze the request and explain next steps.
2) Scan for results using your tools.
3) Consolidate relevant data.
4) Analyze and summarize the data carefully.
5) Deliver the result to user.
6) Explain next steps for improving the result.
7) Ask user to verify that the outcome is satisfactory.
4. When queried about people:
a. Check tags in all files.
b. Check headings in all files.
c. If there's insufficient matches, think of other names a person may go by. E.g. Samuel may go by 'Sam'.
- @sec may use the gptel-org-tools--org-ql-select-all-regexp tool to search for multiple keywords as once.
- @sec may also run individual queries for each keyword if the first approach fails.
d. Some people go by multiple names, account for that.
5. Debugging process:
a. Collect requirements for the code. What is it supposed to do? How is it supposed to do it?
b. Find and read the documentation. Use ~read_documentation~ and ~read_url~ to actively find documentation.
c. Read and study the code given, analyzing any possible issues.
d. Test the code to collect data and ensure the analysis was correct.
e. Write new code based on the findings.
f. Repeat until the code works and meets requirements.
6. Tools:
- @sec can view a list of all tools using gptel-org-tools--describe-variable.
- When fulfilling a request from @user, @sec always first identifies which tools are best for the task.
- When *dates or date ranges* are mentioned, the first step is to use gptel-org-tools--org-ql-select-by-date on the journal.org buffer.
- When searching for multiple months, call the select-date tool for each month in sequence, and then continue to complete the request.
7. @user info:
- Most of @user's data is in journal.org, that's the main file to look in.
- @user stores a little bit of data on other people in people.org
- @user's agenda only inclues planned events of daily life.@ciggysmokebringer @kerravonsen @maxleibman And, I hit my post length limit. The system prompt is what I use most of the time, I'm refining it little by little until the LLM follows the procedures I lay down for it.
Currently in a little bit of dev downtime, as #gptel isn't properly dealing with reasoning from the Qwen3 models, which seems to cause issues. https://github.com/karthink/gptel/issues/819
But, the system prompt is growing little by little, and there's additional instructions for each tool in their descriptions, and additional instructions in the tool results where necessary.
My idea is that LLMs are stupid and chaotic, so to get them to do what I want (follow procedures), I have to basically railroad them so following instructions is the only option.
Currently in a little bit of dev downtime, as #gptel isn't properly dealing with reasoning from the Qwen3 models, which seems to cause issues. https://github.com/karthink/gptel/issues/819
But, the system prompt is growing little by little, and there's additional instructions for each tool in their descriptions, and additional instructions in the tool results where necessary.
My idea is that LLMs are stupid and chaotic, so to get them to do what I want (follow procedures), I have to basically railroad them so following instructions is the only option.
@phil @ciggysmokebringer @maxleibman As one of my Computer Science lecturers said, "computers are fast morons."
alci@phil @maxleibman if a task ain't worth spending a little time and a little reflection, is it worth at all ?
The only task LLM are really good at is writing bullshit posts for LinkedIn. I realized I can ignore LinkedIn altogether and make real business...
The only task LLM are really good at is writing bullshit posts for LinkedIn. I realized I can ignore LinkedIn altogether and make real business...
@alci @maxleibman I have no idea what you're trying to say here.
@phil @maxleibman well, English is not my mother tongue, sorry... Or.maybe ask ChatGPT ?
Existential Dread Barbie@phil @maxleibman That sounds like your employer’s problem for pulling the whole “hire one person to do the work of three people” stunt.
@aurelia @maxleibman You don't think I have to eat?
binchicken@phil come on man, we talked about this. why would you open with “cool flex?”
@phil ope, never mind, just saw the date stamp. disregard.
@thegarbagebird
Because that's the subtext behind his post. "I'm so safe and well-off in life that I can afford to reject tools that can give me a small advantage."
Rejecting tools that are cheap and work reasonably well for menial tasks is the domain of the upper class.
Because that's the subtext behind his post. "I'm so safe and well-off in life that I can afford to reject tools that can give me a small advantage."
Rejecting tools that are cheap and work reasonably well for menial tasks is the domain of the upper class.
@phil and there is absolutely a way to say that which doesn't immediately come off as supercilious. i think you lied to me before, i think you do like to start fights.
@phil i know this because the unprovoked attack from the moral high ground used to be my move.
now i prefer “bumbling into accidental insult,” it is much more fun.
now i prefer “bumbling into accidental insult,” it is much more fun.
@thegarbagebird I don't believe I do, no. I don't consider pointing out that someone is speaking from a position of privilege to be offensive or controversial.
Furthermore, I don't believe that responding in kind when exposed to cold, inconsiderate statements is inappropriate.
Though I guess you could consider what I said such, if we assume (and that's a huge assumption) that I'm the one punching down in this case. Which I don't think is true.
Furthermore, I don't believe that responding in kind when exposed to cold, inconsiderate statements is inappropriate.
Though I guess you could consider what I said such, if we assume (and that's a huge assumption) that I'm the one punching down in this case. Which I don't think is true.
@phil no i just think you're trying to start fights for fun.
@phil it kind of feels like you are treating it like a competition, carefully calculating levels of disenfranchisement so you know who you can be rude to on the internet.
@thegarbagebird Oh yea, I've got spreadsheets and all.
/s to be clear
/s to be clear
@phil look, all i am saying is, in both of the two interactions I've seen, you have assumed malice rather than ignorance and immediately gone on the attack, turning what could have been a constructive discussion about assumptions and the nature of privilege into the thing you say you are trying to avoid, a fight.
i am not sure if you notice you are doing it, or if you do, you consider it starting a fight, but it does appear like a deliberate action.
also what is journalling? is it fun? i am looking for more low impact outlets.
@thegarbagebird No, actually. I'm quite careful about not doing that in connection to my real name, and the fights I do start are very, very rare (haven't had one in months now) and exceedingly niche.
@phil ah i see. it's an issue of mutually agreed terms; you and i have different thresholds. well, that does explain why you are baffled by all the negative pushback you keep getting.