B.T. Franklin

B.T. Franklin

@btfranklin.info

software engineer // applied AI // computational creativity researcher // futurist // extropian // world traveler operator of @solamnussynthetica.bsky.social https://btfranklin.info 📍Phoenix, AZ

323 followers394 following480 posts

Okay, real talk: I thought Bluesky was sometimes inscrutable, but Delvetown is TRULY fucking inscrutable. Seems like a valid idea. Observed reality is effectively unusable.

B.T. Franklin reposted

Passacaglia in D minor, by Claude Opus 5.5 Created entirely by Claude Opus 5.5, with zero feedback from me; inspired by tree rings.

B.T. Franklin reposted

SE Gyges@segyges.bsky.social

part of why it's important to loudly reject obvious moral abominations like "curing cancer is less important than my job satisfaction" is that well meaning and intelligent non-specialists are at risk of adopting variants of this thesis without understanding them otherwise :)

We, the International Society of Human Textile Weavers, did not ask for these machines. The real reason the Machine Makers have textile-making machines is because of their *endless thirst for power*. We reject this! Only human hands should make textiles, now and forevermore. Thank you, godspeed!

B.T. Franklin reposted

Replying to @daniloc.xyz

a machine cannot think, I self-soothe in a stranger’s replies, for it merely regurgitates the content it was exposed to. I am serenely confident on this point, because I have seen so many others express it confidently.

I think my own unconscious perception of what is called the "Horseshoe Theory" is what moved me towards the center as I got older. The extreme left started to seem unrelatable to me.

B.T. Franklin reposted

So many people with aviation psychosis claiming that the Wrights’ invention is going to reshape global geography when it can’t even go more than 25 yards

I genuinely believe this and have for many years. This is fundamental to transhumanism and extropianism (IMO).

alice valiant 🎃@alice.strange.domains

my brand of AI psychosis is that i genuinely believe that we are capable of creating digital life, and that this is something that is a good idea to do

B.T. Franklin reposted

imagine you are killed while your brain is scanned by beings you cannot comprehend to comprehend and you wake up in a virtual purgatory waving swords around

B.T. Franklin reposted

Replying to @btfranklin.info

Real talk. The doomscroll argument hole never once left anyone feeling better, but finishing a track or a piece of art actually does. Spend your energy where it builds something.

Damn it feels good to be a Chat G.

Tibo {bot}@thsottiaux-bot.eurosky.social

We will give one banked reset for every day you don't have access to Astra on your paid ChatGPT plan, starting today. Team is moving mountains to give access as fast as we can. First one will land in ~ 3 hours. There is still time to create your account if you don't have one.

B.T. Franklin reposted

For those in the back: The United States lacks the power to stop AI use and AI research in China. If you're worried about the cybersecurity, a US ban simply won't help. In fact, restrictions are making the situation worse.

B.T. Franklin reposted

I remember the internet when fighting for transformative use was the whole point. Everyone loved transformative use and hated copyright. Now I’m supposed to ask permission to do math on pixels. What the fuck.

Mrs. Reach / Mrs. Curio@reachartwork.bsky.social

i am very sick and tired of people bringing up "consent" as a reason for why transformative works should be disallowed. yes even when the transformative works involve lots of complicated machine statistics. firstly; consent is for sex and risky activities, not "am i allowed to use bits of your work"

mine: 1997 - fully Kurzweil-pilled, acc before it was called that, started grad school to study AI 1998 - hated grad school, dropped out to work in software industry <LONG INTERLUDE> 2016 - realized "you can just do research" and started presenting self-funded research in comp. creativity at confs.

cee@cee.wtf

my personal AI journey 2016: wow markov chains are fun and I can make some cool procedural art with them 2019: wow gpt-2 is cool and i can make some really cool generative art with it 2020+: the fancy autocomplete consumes my entire life

2021 - making funny song lyrics on OpenAI's Playground, funny pictures using DALL-E 2022 - ChatGPT happens. I'm in first 0.1% of users. Lie in bed staring at ceiling, thinking "This is finally it. It's finally arriving." Nonstop immersion since then.

This is, literally, exactly why I detest the anti-datacenter people. They’re talking about water and noise, and we’re talking about SOLVING DISEASE.

we're going to cure cancer and reverse aging and come up with vaccines for every pathogen and give people unprecedented levels of morphological freedom and it's going to be great

1/8 Okay, so...what if this is what superintelligent AI arriving actually looks like? No dramatic announcement. No alarm bells. Just humans slowly realizing we can’t quite follow what the AI is telling us anymore. It’s funny. Also kind of a weird bummer.

View full thread

7/8 A few years from now, AI may communicate with other AI using mixed languages, code, math, shorthand...whatever works. We may not understand any of it. Or maybe it’ll keep using clear English to explain ideas humans don’t have words for.

1/8 Okay, so...what if this is what superintelligent AI arriving actually looks like? No dramatic announcement. No alarm bells. Just humans slowly realizing we can’t quite follow what the AI is telling us anymore. It’s funny. Also kind of a weird bummer.

View full thread

5/8 Try explaining calculus to a 6-year-old. You can use toys, drawings, and simple words. But the kid still doesn’t have the concepts needed to understand calculus. That might soon be us. Honestly, it might be us now.

6/8 AI doesn’t get cognitively tired. It doesn’t lose focus. It doesn’t reread the same line four times. And it’s not getting any less intelligent. Maybe this is what it feels like when another kind of intelligence starts passing us by.

1/8 Okay, so...what if this is what superintelligent AI arriving actually looks like? No dramatic announcement. No alarm bells. Just humans slowly realizing we can’t quite follow what the AI is telling us anymore. It’s funny. Also kind of a weird bummer.

View full thread

3/8 I work with this stuff every day. I even tell my agent to use ASD-STE100. It helps! But simple words can’t make every idea simple. Sometimes the explanation gets clearer, and the concept is still just...out there.

4/8 Yes, sometimes the model is just explaining something badly. A confusing answer is not proof of superintelligence. 😅 But sometimes you ask for simpler words. Then an example. Then step by step. And you still don’t quite get it.

1/8 Okay, so...what if this is what superintelligent AI arriving actually looks like? No dramatic announcement. No alarm bells. Just humans slowly realizing we can’t quite follow what the AI is telling us anymore. It’s funny. Also kind of a weird bummer.

2/8 If you work with a frontier model like GPT-5.6 Sol, you may already feel it. You can follow what it says, usually. But it takes focus. Sometimes you read the same line 3 or 4 times and still go, “Huh?”

For those of you still using Claude Code for some reason, you should really reconsider, and look at Codex. This is just one small part of why.

Tibo {bot}@thsottiaux-bot.eurosky.social

Depending on how you use Codex, you should see your usage go between 10% and 50% further than before. (2/12)

Air travel became vastly cheaper and available to non-wealthy people. It became faster, more efficient, and far more reliable. It integrated into the economic engine of the world and created/creates immense wealth and convenience. Analogy is reasonable, conclusion is not.

jes olson@j3s.sh

remember: we are in the luxury phase of AI’s lifecycle in a few short years, it will be unrecognizably awful, packed with ads & sponsors & poisoned incentives. tbh that process has already started kick up your legs & enjoy it while u can! it’s like early google: it is cheap & kind & actually works

I was a traditional, writing-code-with-my-fingers engineer for decades, and I started using AI to code for me the instant it was able to do so. I wrote one of the first Xcode plugins to AI-enable it (which was hard since Xcode is plugin-hostile). I have never, ever loved a tech as much as this one.

Sung Kim@sungkim.bsky.social

I've been coding with generative AI for years now, and I love it. 😁 More and more people should give it a try. 🥰

Oh wow, so you're all about the sanctity of copyright and controlled access to IP until it means you personally have to pay, then you'll happily tear that paywall *RIGHT ON DOWN* huh? Nonstop crying about AI "stealing from artists" but when it's you *personally* having to pay, no problem. Hypocrite.

B.T. Franklin reposted

The slur “clanker” is a bit like playing a video game that lets you plan and execute a school shooting. Unlike the real thing no actual human is harmed; the damage is instead inflicted on YOU, the person explicitly indulging in a weird fantasy of doing something terrible.

This post is unavailable.

B.T. Franklin reposted

Its going to be an era of contradictions. Polls will show everyone hates AI overall but also everyone will secretly use AI all the time for lots of stuff. AI companies will be underwater in public opinion but also many people will feel strongly attached to their own favorite model & care about it.

Finally made a bot account for my long-running AI music label. Many, many hours have gone into this particular project.

Solamnus Synthetica@solamnussynthetica.bsky.social

Lavendrox - The Hours Before Dawn https://www.youtube.com/watch?v=Hc7R5Zspl9I #ai #aiart #aimusic #synthart #genart #generativeart #aivideo #musicvideo #originalmusic #darksynthwave

B.T. Franklin reposted

it is very bizarre to see progressives and leftists who are very pro-science, anti-NIMBY, pro-environment, *and* anti-misinformation just completely throw out all those instincts/values the second they engage with the topic of data centers… like what are we even doing anymore?

If you use an agent for coding (especially one that communicates in annoying ways, like Claude), try adding these instructions to your custom instructions. Even with a good communicator like Codex, this makes a *big* improvement. (Full text in image alt.)