thethunderwolf
@thethunderwolf@lemmy.dbzer0.com
A random nerd on the internet
Pronouns: they/them
- Comment on I don't know the reason why. 9 hours ago:
don’t default to “he”, say “they”
- Comment on I don't know the reason why. 9 hours ago:
US defaultism
- Comment on Lemmings, please give us your info dump. 11 hours ago:
I’ve heard of this scenario as an example of why not to put your face on the internet. Now with AI it’s actually happening.
- Comment on Lemmings, please give us your info dump. 11 hours ago:
capacity for sentience
Aren’t most animals sentient? Wiktionary says that it means “Experiencing sensation, thought, or feeling”. Even simple animals like flies are capable of “sensation”, and most complex animals experience all three. I would say “sapient” (“possessing intelligence or a high degree of self-awareness”) in this context.
You’re the scientist, what is the scientifically correct/used/accepted meaning of “sentient”? Or has it been found that most animals do not, or at least do not consciously in the way humans do, sense, think, or feel? Correct me if any of what I’m saying is wrong.
- Comment on Better be a sex offender registry on those damn space ships 1 day ago:
“The government”
the one IANA regards as THE “government”
- Comment on Homeland Security has reportedly sent out hundreds of subpoenas to identify ICE critics online 1 week ago:
“Woke liberal commie antifa terrorism”
- Comment on Homeland Security has reportedly sent out hundreds of subpoenas to identify ICE critics online 1 week ago:
ICE engines function well but are quite suboptimal, they pollute the air
Developing a good replacement (such as electric cars) is important
- Comment on Homeland Security has reportedly sent out hundreds of subpoenas to identify ICE critics online 1 week ago:
Welcome to Lemmy!
- Comment on Google criticizes Europe's plan to adopt free software 1 week ago:
competition is bad for competition 🤨
- Comment on 1 week ago:
I understand why there are paid features; I just personally intend to wait until federation drops so I can join an instance that offers Plutonium for free.
- Comment on 1 week ago:
And if you selfhost you can get the stuff for free
- Comment on 1 week ago:
Movim seems less like a Discord alternative and more like a Whatsapp alternative
- Comment on 1 week ago:
Paid features? Ew.
I’ll be waiting until federation rolls out. Someone will definitely create a fork that gives you the paid features for free
- Comment on cant take it anymore 1 week ago:
You weren’t yet
Soon you will have been
- Comment on Ars Technica makes up quotes from Matplotlib maintainer("An AI Agent Published a Hit Piece on Me"); pulls story 1 week ago:
[citation needed]
- Comment on Ars Technica makes up quotes from Matplotlib maintainer("An AI Agent Published a Hit Piece on Me"); pulls story 1 week ago:
what?? AI is not conscious, marketing just says that with no understanding of the maths and no legal obligation to tell the truth.
Here’s how LLMs work:
The basic premise is like an autocomplete: It creates a response word by word (not literally using words, but “tokens” which are mostly words but sometimes other things such as “begin/end codeblock” or “end of response”). The program is a guessing engine that guesses the next token repeatedly. The autocomplete on your phone is different in that it merely guesses which word follows the previous word. An LLM guesses what the next word after the entire conversation (not always entire: conversation history may be truncated due to limited processing power) is.
The “training data” is used as a model of what the probabilities are of tokens following other tokens. But you can’t store, for every token, how likely it is to follow every single possible combination of 1 to <big number like 65536, depends on which LLM> previous tokens. So that’s what “neural networks” are for.
Neural networks are networks of mathematical “neurons”. Neurons take one or more inputs from other neurons, apply a mathematical transformation to them, and output the number into one or more further neurons. At the beginning of the network are non-neurons that input the raw data into the neurons, and at the end are non-neurons that take the network’s output and use it. The network is “trained” by making small adjustments to the maths of various neurons and finding the arrangement with the best results. Neural networks are very difficult to see into or debug because the mathematical nature of the system makes it pretty unclear what a given neuron does. The use of these networks in LLMs is as a way to (quite accurately) guess the probabilities on the fly without having to obtain and store training data for every single possibility.
I don’t know much more than this, I just happen to have read a good article about how LLMs work. (Will edit the link into this post soon, as it was texted to me and I’m on PC rn)
- Comment on "I am going to punch you" WHAT A BOSS! 1 week ago:
*Public disservice
- Comment on Discord will restrict your account next month unless you scan ID or face 1 week ago:
Bluesky is poorly federated
Lemmy is excellently federated:
- Its users are split across many instances with no majority
- Lemmy, PieFed, and other instance software exists, preventing anyone from seizing control of the network software
- Comment on Discord will restrict your account next month unless you scan ID or face 1 week ago:
Does anyone seriously believe there’s an indie third party social media site immune to a multi-million (possibly even billion) dollar buyout offer? One that won’t go the way of TikTok and be legislatively forced into sale?
federated platforms
- Comment on The Department of Homeland Security Is Demanding That Google Turn Over Information About Random Critics 1 week ago:
I was merely listing non-Google AI models. The actual recommendation that I gave is “I’d recommend not using AI”, but because some people find AI useful and like it, I mentioned the popular other options and also “anything else”.
I don’t use AI, so I don’t have an ethical FOSS model to recommend. If someone else does, I’d be curious to know.
- Comment on The Department of Homeland Security Is Demanding That Google Turn Over Information About Random Critics 1 week ago:
I’m not sure
- Comment on 'What a great way to kill your community': Discord users are furious about its new age verification checks — and are now hunting for alternatives 1 week ago:
sadly people will forget about this next week
- Comment on 'What a great way to kill your community': Discord users are furious about its new age verification checks — and are now hunting for alternatives 1 week ago:
and Australia
- Comment on The Department of Homeland Security Is Demanding That Google Turn Over Information About Random Critics 1 week ago:
- Gemini: I’d recommend not using AI, but if you want to, then use Deepseek, Claude, ChatGPT, or anything else
- Google Search: DuckDuckGo, SearXNG, Kagi, Mojeek, Qwant, Startpage, Marginalia
- YouTube: use invidious.io for watching YouTube videos (it also gets rid of ads), and use PeerTube (video platform federated like Lemmy) and watch videos, and maybe post some if you have something to post about
You can also search the web for lots of deGoogling advice and guides and for alternatives to Google services
- Comment on The Department of Homeland Security Is Demanding That Google Turn Over Information About Random Critics 1 week ago:
Or real non-Android-based mobile Linux?
It’s about time for the year of the Linux phone
- Comment on Western Digital details 14-platter 3.5-inch HAMR HDD designs with 140 TB and beyond 1 week ago:
keep the OS partition like 4tb and make a separate partition for the pirated movies
- Comment on The new Microsoft copilot key is impossible to properly remap. 2 weeks ago:
Apple is also a patent troll
All megacorps are
Patents need to be abolished
- Comment on Top of the world, ma 2 weeks ago:
why
- Comment on Top of the world, ma 2 weeks ago:
bot account 🤨
first interaction by account 🤨 - Comment on no shit 2 weeks ago:
It