Anthony

I'm Anthony and I'm an organic computer scientist and a Luddite. I believe that unsound research lends itself to social harms.

website: https://bucci.onl/about.html

Public Key

npub1j4dn7g4j0eu0q69xvunf2uqk5f7npkjvgxtnrep6yezyv9h6nf0snnt0rs

NIP-05 Address

abucci_at_buc.ci@momostr.pink

Profile Code

nprofile1qqsf2kely2e8u78sdznxwf54wqt2ylfsmfxyr9e3usazv3zxzmaf5hcpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dswl60p4

Publishing to

nostr.mom

relay.damus.io

nos.lol

Author Public Key

npub1j4dn7g4j0eu0q69xvunf2uqk5f7npkjvgxtnrep6yezyv9h6nf0snnt0rs

Show more details

Last Notes

2026-03-16 23:32:25 UTC

- reply

Seems it's time for an LLM-free Linux distribution, perhaps forking useful projects before they went off the deep end.

2026-01-23 17:42:58 UTC

- reply

This is not quite an answer, but to the extent that corporations are the main purchasers of image generators, the promise seems to be similar to what LLMs promise. Confine human endeavor to a small, controllable world that seems "creative", but is incapable of generating a true challenge to the powerful. Here I am using Leonard Savage's small vs. large world distinction. The former is like living in a chess game according to the rules of chess; the latter is like living in the real world, where you can knock over the board, refuse to play, draw a smiley face on the Queen, etc etc etc. If chess grandmasters convince people that all of society's concerns should be decided over a chess board according to the rules of the game then they've locked in their power because there is no rule of chess that allows you to suspend the rules of chess. That's how I've been seeing the value of AI (to the powerful) lately.

2025-12-19 15:14:07 UTC

- reply

Is it always off by default? Are all of the configuration options it covers off by default and stay off even if I turn the kill switch back to on?
Are all the options listed here controlled by the kill switch? https://buc.ci/abucci/p/1763845084.289082
@npub1lyp…a5h0 @npub1dpj…e0jl

2025-12-19 15:10:41 UTC

- reply

The user experience sucks because I don't want AI anywhere near my computer, and I don't want to have to put in work a *web browser* to ensure this. By adding these features you've introduced more friction in the form of a configuration tax each and every time I update the browser.
@npub1lyp…a5h0 @npub1dpj…e0jl

2025-09-16 20:45:18 UTC

> Put plainly, you should feel bad for using AI.
Yes.
From https://www.nplusonemag.com/issue-51/the-intellectual-situation/large-language-muddle/
#AI #GenAI #GenerativeAI #LLM #Claude #Copilot #Gemini #GPT #ChatGPT #tech #dev #science #research #writing

2025-06-06 17:42:34 UTC

- reply

Besides the nanny state I worry about elite panic ( https://en.wikipedia.org/wiki/Elite_panic ). It's hard to be sure whether the "information" you're being told during an adverse event is factually accurate, or whether it reflects elite panic. There's probably a bit of both going on most times.

2025-06-04 15:28:31 UTC

- reply

They quite literally are not. Though the story goes that Hopfield used to have students in his classes manually implement Hopfield networks to demonstrate how they worked viscerally.

2025-06-03 16:24:26 UTC

- reply

My son asked me what this book was, so I told him, and now he refers to him as "Max Forkheimer".
I opted to start at the beginning so I'm partway into *The Concept of Man*. I'll let you know if I have any deep thoughts!

2025-04-29 19:48:48 UTC

- reply

Sudden flashback to those "this is your brain on drugs" TV commercials! (https://youtu.be/GOnENVylxPI)

2025-04-29 19:31:53 UTC

- reply

But that rhetoric always was and always will be hypocritical and self serving. Ayn Rand collected Social Security and Medicare benefits, etc. etc.

2025-04-28 17:10:15 UTC

- reply

Just throwing in my 2 cents that I don't think that data poisoning is as effective as forcing companies to pay an acquisition cost for good text. Doubtless they have a QA pipeline, one that includes low-paid workers overseas, that can filter out generated text pretty effectively and at low cost. They'll have a much harder time avoiding tarpits, which could really slow down scraping if deployed at a large enough scale. As @npub1zt7…uwkx suggested, having the tarpits be as fast as possible and as diverse as possible furthers that aim.
I was also thinking a model like what signature-based malware or adware detectors use would be cool too. Whenever one tarpit discovers a new technique for snaring bots or keeping bots snared longer, it pushes that innovation to a database where other tarpits can download it and adjust their own strategy, allowing tarpits to collectively adapt faster than the bots can.
Anyway thanks for the good work!

2025-04-28 16:51:18 UTC

- reply

> I don't think I even consider this unethical based on how the research was conducted by humans and not automated.
You should start considering it unethical, because it is. The level of automation is not what determines ethics in human subjects trials; nor does it matter whether the unethical behavior is conducted on a mass scale by others. The "reseachers" knowingly broke the rules of the subreddit, and the moderators themselves, who were unaware of the "experiment", came down strongly against what was done.
All in all a really lousy thing to have done.

2025-04-28 14:25:18 UTC

My half-baked deep thought of the weekend:
Arrow’s Impossibility Theorem should be renamed Arrow's Context-Sensitivity Theorem, and re-interpreted as saying a social choice function that neglects context leads to dictators.
I say this because the axiom of independence from irrelevant alternatives--one of the assumptions behind the theorem--states that a social choice function should be such that the relationship between A and B is not changed once a new alternative C is introduced. Unpacked, this means the choice function should be insensitive from any context C might bring with it.
Arrow's theorem essentially says that a social choice function satisfying this and a couple other axioms leads to dictators (meaning, one individual's preferences dictate the social choice function's preferences, overruling everyone else involved in the choice who might disagree). Hence the re-interpretation: neglecting context in social choice leads to dictators.
#economics #SocialWelfare #SocialChoice #WelfareEconomics #DecisionTheory #ArrowsTheorem

2025-04-24 16:56:48 UTC

- reply

Drew DeVault's post about the impact AI crawlers were having on Source Hut was eye opening: https://drewdevault.com/2025/03/17/2025-03-17-Stop-externalizing-your-costs-on-me.html
You could almost understand the crawling. Spamming issues seems nihilistic.

2025-04-08 15:50:06 UTC

It's not a dire wolf. That's obvious and requires no further deliberation. The priests of the church of techno-optimism are loud and platformed but you don't have to take their sermons at face value.

2025-04-08 15:48:00 UTC

- reply

My take on this is that it's setting the stage for climate-change-denying religious techno-optimism of the form "Don't worry if species go extinct. We have the technology to bring them back."

2025-04-08 11:30:10 UTC

- reply

It seems the message they are trying to get people to accept is "don't worry about species going extinct. We have technology that can bring them back." Fairly typical incantation of the technooptimist religion if you look at it that way.

2025-04-06 16:57:42 UTC

Spent about two hours working in the woods today. This property had been neglected for years and it shows. We've been removing briars, brush, and vines that are choking trees, clearing paths, and taking down sick or dead trees for a couple years now but there's so much left to do. Eventually we'll have the ancient rusted car that's back there removed (!). Anyway today we cut another path from the yard back into the woods to a clearing we made last summer. It's already a nice place to sit and relax.

2025-03-07 16:57:40 UTC

- reply

Speaking of blocking public access to help: https://buc.ci/abucci/p/1741366568.271730

2025-03-07 16:27:49 UTC

- reply

Success! Thank you.

2025-03-07 15:08:08 UTC

- reply

Hi Yogi. I just tried to register for this, and was given an error: This meeting is for authorized registrants only. Please enter another email address
It seems the registration link might not be configured to be open to everyone right now.

2025-03-06 20:47:01 UTC

- reply

The first emails were like:
> I came across Bucci online and wanted to reach out because we're working with other computer scientists that are offering computer science services.
> ‍ I’m currently working with a few computer scientists on campaigns that are getting their sales teams meetings with warm leads every month. We’re targeting people who need computer science services using personalized email sequences. ‍
Nothing I like better than getting my sales team warm leads targeting people who need computer science services.

2025-03-06 20:25:31 UTC

- reply

Well well well, if it isn't my old friend Jesse Pines!
https://buc.ci/abucci/s/d2b5ca33bd970f64a6301fa75ae2eb22.png

2025-03-06 17:52:22 UTC

- reply

Thank you. It sounds like a reasonable way to go. In my case I'm using Sublime Text, and it picks up which JVM to use from the user environment.

2025-03-06 16:30:07 UTC

- reply

Yeah, I'm not trying to be critical or anything, just trying to get my own head straight and understand where you're coming from better. Talking about it definitely helps. @npub1hyn…0yth

2025-03-06 16:16:24 UTC

Nowadays programming in a programming language I don't use daily seems to always require an upgrade cascade of editors, tools, plugins, dependencies, libraries, my DNA, ??? I put some effort into keeping my environment static but all it takes is one autoupgrading thing I missed to kick off one of these cascades, and it feels like whack-a-mole trying to find and lock down every possible cause. This time it looks like a newer version of scala metals might have stopped supporting Java 11 and somehow got updated without my knowledge (maybe? I'm guessing).
P.S. This is not an invitation to post critiques about any of these technologies or recommendations about what I should be doing instead.
#scala #dev #tech #SoftwareDevelopment #coding #programming

2025-03-06 15:38:54 UTC

- reply

You really believe I'm eliding the downsides in one direction and exaggerating the other? That's the core of the disagreement then, seems to me.
Jumping the gun can lead to needless loss of life, waste of resources that are then not available when the situation devolves further and you need them, etc. If you're going to suggest I might be eliding the downsides in one direction, don't turn around and do the same thing in the other! It's worth reading the Wikipedia article, and the linked FEMA articles, on emergency management for more on this.

2025-03-06 15:24:48 UTC

- reply

If someone is planning to invade your house, you can: call a friend for help; call the authorities; let your dog off the leash; prepare some kind of defensive weapon; hide your valuables; lock your safe if you have one; talk to everyone in the house about what might be coming and what you'll do; etc etc etc. You have a chance of preventing the invasion from ever happening, or minimizing the effect if it does. You're not physically injured.
If someone is in the *midst* of invading your home, your range of action collapses. You are now in the mode of managing the situation. You cannot prevent the invasion, and may not be able to prevent its worst effects anymore. You're at immediate risk of loss or injury.
If someone has already invaded your home, you are in recovery mode. You may have already lost things, perhaps permanently. You may be physically injured.
Likewise if a tornado is coming you can go down in the basement or brace yourself in a sturdy doorframe. If you have enough warning you might even be able to secure doors and windows. Those actions are pointless if the tornado has already passed; at that point you're tending to whatever broke.
Standing at the top of the ski hill thinking about bombing down it is very different from actually bombing down the hill, which is very different from being wrapped around a tree.
Etc.
These are risk and emergency management distinctions: [prevention, mitigation, preparedness, response and recovery](https://en.wikipedia.org/wiki/Emergency_management#Phases_and_personal_activities ). Jumping the gun--going into "response" mode too soon for instance--has numerous downsides.
@npub1hyn…0yth

2025-03-06 14:45:45 UTC

- reply

There's a big difference between rhetoric or planning and action. It matters. It doesn't mean we ignore the threats, but we also don't take them as if they've already been enacted. We don't help ourselves by collapsing this distinction. We don't gain any power, or insight, or strategic guidance.
@npub1hyn…0yth

2025-03-06 14:28:16 UTC

- reply

With respect, I think that's going too far. MTG isn't storming into Congress with weapons and goons and dragging AOC away. She's acting (to some degree) like a member of Congress. They're using the levers of power that are there. Yes, they're doing damage and yes, they're remaking them. But they're not throwing them off or smashing them.
And I believe they can't do that, at least not right now, because the US economy would collapse and some of the basis of their current power would dissolve. The US needs trading partners, and USD dominance is a key part of that. One thing worth watching is whether US GDP really does contract by 2.8% as the Atlanta Fed is estimating right now, and how the Trump administration reacts to that.
US citizens are going to Republican town halls and shouting at their Congresspeople. People don't do that when fascists take power. The Congresspeople are not responding by lining up armed thugs in their town halls and dragging off anyone who protests. They're responding by not holding town halls. This suggests to me that they're afraid of their constituents, not that they're fascists confident in their power.
There are street protests happening. They would not be happening if we were really in fascism territory because the analog of brownshirts would be shutting them down. Being harassed or arrested by police is not the same thing.
So, I have to disagree with you. There are important indicators we need to watch, and who knows how those will change over time. But for now I think the indicators are telling us the distinction does matter, and recognizing that aims our attention in what I think are important directions.
@npub1hyn…0yth

2025-03-06 14:13:09 UTC

- reply

The scolding really gets under my skin. It doesn't seem as common anymore but I thought "Vichy Democrat" was a pretty good burn.
@npub1hyn…0yth

2025-03-06 14:07:44 UTC

- reply

I grew up in similar places, and I've also read a bit of history (e.g. of the KKK). I don't dispute that these views are dangerous and horrifying, nor that Americans are capable of extreme violence. However, I think it's important to be careful with words. I think Richard Seymour's book *Disaster Nationalism*, and his various talks and [articles](https://www.theguardian.com/environment/2024/oct/30/richard-seymour-on-far-right-environmental-crisis-disaster-nationalism ) on this topic, are important. He characterizes it more as a "proto" fascism. It's lacking certain features that actualized fascist regimes in the past had. Anyone who cares about doing something about all this would do well to pay attention to the distinctions.
From the article I linked (bold added by me):
> Comparing the success of the far right in India, Brazil and the US (among other places), Seymour argues that most explanations for their rise are insufficient. What we’re seeing is “too consistent over time and too global, to be explained by local factors such as the backlash of a fading white supremacy, or Russian troll farms, or ‘bad actors’ spreading disinformation,” he writes. > **These movements also don’t have the hallmarks of historical fascism**> . “Their immediate objective is not the overthrow of electoral democracy,” Seymour observes, but “a constitutional rupture breaking with all humane and ‘woke’ constraints on the exercise of power.” While the old establishment decomposes, the far right conjures up apocalyptic images – “the great replacement”, “Islamisation”, “Chinese-style communism” – to animate potential supporters. This is not yet a distinct form of fascism; instead, it is what Seymour calls “disaster nationalism”.
@npub1hyn…0yth

2025-03-06 13:58:00 UTC

- reply

I'm not sure what you mean by saying that Nazis ushered in Trump. For my part, I find the "Nazi" rhetoric at best unhelpful, and at worst actively harmful. It misdiagnoses the situation, it reinforces a polarization within the electorate that isn't real and benefits the powerful (they *always* divide and conquer us; why do their work for them?), it offers a cartoon picture of what's happening ("punch a Nazi!" posters) with no strategic guidance, and it's hypocritical (*).
@npub1qhh…sz74
(*) If anyone *really really* believed these people are Nazis why aren't they neutralizing them immediately? I think it's because they *don't* believe this, it's just rhetoric they've gotten into the bad habit of using. MAGA-infected folks call liberal-minded people fascists and Nazis too. We're really just going to shout these words at each other over and over? Till what happens, exactly?

2025-03-06 13:26:54 UTC

- reply

Speaking as someone who lays a sizable responsibility at the feet of the Democratic party, and who firmly believes they earned that and more, I don't understand what you mean by saying blaming that party is "pathological".