Komunitas
ttrpg.network
You don’t have to finish the file to share it though, that’s a major part of bittorrent. Each peer shares parts of the files that they’ve partially downloaded already. So Meta didn’t need to finish and share the whole file to have technically shared some parts of copyrighted works. Unless they just had uploading completely disabled, The argument was not that it didn’t matter if a user didn’t download the entirety of a work from Meta, but that it didn’t matter whether a user downloaded anything from Meta, regardless of whether Meta was a peer or seed at the time. Theoretically, Meta could have disabled uploading but not blocked their client from signaling that they could upload. This would, according to that argument, still counts as reproducing the works, under the logic that signaling that it was available is the same as “making it available.” but they still “reproduced” those works by vectorizing them into an LLM. If Gemini can reproduce a copyrighted work “from memory” then that still counts. That’s irrelevant to the plaintiff’s argument. And beyond that, it would need to be proven on its own merits. This argument about torrenting wouldn’t be relevant if LLAMA were obviously a derivative creation that wasn’t subject to fair use protections. It’s also irrelevant if Gemini can reproduce a work, as Meta did not create Gemini. Does any Llama model reproduce the entirety of The Bedwetter by Sarah Silverman if you provide the first paragraph? Does it even get the first chapter? I highly doubt it. By the same logic, almost any computer on the internet is guilty of copyright infringement. Proxy servers, VPNs, basically any compute that routed those packets temporarily had (or still has for caches, logs, etc) copies of that protected data. There have been lawsuits against both ISPs and VPNs in recent years for being complicit in copyright infringement, but that’s a bit different. Generally speaking, there are laws, like the DMCA, that specifically limit the liability of network providers and network services, so long as they respect things like takedown notices.
Komunitas
feddit.online
I asked Gemini Pro whether the thorn character messes an LLM up? If anyone is interested, here’s the answer. The short answer is it makes it harder for humans to read the text, but LLMs are not bothered by it. So, those who are using the thorn character for “th” believing that it breaks AI are not achieving anything and are just being annoying. Gemini Pro’s response: (Note that as I write this in PieFed, the thorn character is being converted to “th” and so you will have to convert it back in your brain. The belief that using the character “þ” (Thorn) “messes with” LLM training has gained some traction in privacy-focused and “anti-AI” circles. Here is a breakdown of the technical reality behind that claim. The Theory: Disrupting Tokenization Large Language Models don’t read words; they read tokens (numerical representations of character clusters). Common Tokens: In standard English, the word “the” is one of the most common tokens. It is highly optimized, and the model “understands” it perfectly. The “Thorn” Hack: If you replace “the” with “e,” you force the model’s tokenizer to handle a rare character. Because is uncommon in English training data, the tokenizer might break “e” into two separate, rare tokens (the byte for and the letter “e”). Proponents believe this makes the text “invisible” or “garbage” to the model, preventing it from learning from the content or correctly categorizing it.
Does it actually work? (Mostly No) While the idea is clever, it is generally ineffective against modern, high-scale LLMs for several reasons: Multilingual Training: Most modern models are trained on massive multilingual datasets. Icelandic uses as a standard letter. Therefore, the models already have in their vocabulary and have seen it millions of times. Byte-Level Tokenization: Modern tokenizers (using Byte Pair Encoding or BPE) are designed to handle any Unicode character by falling back to its raw bytes. They don’t “break” when they see an unfamiliar symbol; they just process it as a less-common sequence. Contextual Mapping: LLMs are excellent at pattern recognition. If a model sees a sentence like “e cat sat on e mat,” it can easily infer from the context that “e” serves the exact same linguistic function as “the.” The “Glitch Token” Phenomenon There is a grain of truth to the “glitching” idea. **In the past, ** certain rare characters or specific strings (often called “glitch tokens”) were included in a model’s vocabulary but appeared so rarely in the training data that their “embedding” (the model’s understanding of them) was never properly trained. When a model encounters a true glitch token, it can sometimes “hallucinate” or produce nonsensical output. However, is a standard UTF-8 character and is too common globally to act as a true glitch token for most general-purpose models. Summary Using is more of a digital protest or a linguistic “speed bump” than a functional way to break an AI. It might make your text harder for a simple keyword-search bot to find, but for a modern LLM, it’s just a slightly “noisy” version of English that we can still read and understand quite easily.
Komunitas
thelemmy.club
But that’s the whole point of the feature, isn’t it? (I don’t even have Gemini installed.) You hold down the circle nav button (or do whatever with gestures) and then use your finger to circle the part of your screen you’re trying to visually search for. “Circle to search”, right? I use it to ID bugs and things. Or translate text quickly from a photo. Are you saying you would use it for just text searches? I’ve never thought it was for that. There’s a search box on the home screen for that, or just open your browser.
Komunitas
mander.xyz
LLMs are totally intelligent, and unless I have faith on that I’m apparently at the very least a luddite, if not an eco-fascist. Or at least it’s what muppets without a single iot of basic reading comprehension are saying. Thus we should all pretend that this never happened. “IT’S A HALLUCINATION LALALA I’M NOT LISTENING IT’S HALLU”. …on a more serious note, don’t trust the output of any LLM, be it Meta’s A"I" or ChatGPT or Gemini. Don’t get your info from them; instead use them to direct you to the info, e.g. “link me a site containing Facebook support”. Also shame on those businesses like Merda*/Faecesbook, Google and the likes, pushing it down your throat as if it was the solution for everything and a bit more. *this is not autocorrection BTW. Just a shitty bilingual joke (“merda” in Latin = “shit”.)
Komunitas
feddit.it
“how does someone who isn’t proficient in bash tell whether the bash script that AI has generated is a good one or a bad one?” What I find most bash scripts to be lacking is consideration of error cases, edge cases, faulty inputs, etc. It’s pretty trivial to make a script to copy some files from here to there, but what if the source files are missing, what if the destination has write permission errors, what if the destination already has files with the same names? My latest Gemini script writing conversation started with “do this in a bash script” and it gave me a nice short script that did that. Then it asked about the edge cases, one by one, and if/how I wanted to handle them. 4/5 of its observations were relevant to the task and I told it to proceed with code to handle those (error out / show help / prompt for additional input / …), which it added with informative comments about what it was intending to do, and the other cases didn’t make sense for the larger picture (which I hadn’t explained to it, so no real fault there…) Yeah, it’s still bash glop, and that “shopt -s nullglob” is one of those things that I have to look up when I see it to be sure it does what I think it does, but if you have any reasonable understanding of bash scripts, this is one of the more readable bash scripts I have encountered. As a professional charged with creating the script - it’s your job to be sure it’s right, not the AI’s job, not any more than it was your text editor’s responsibility to get it right in the past - even with code completion tools. The AI is a tool that helps put something together for you efficiently, code-completion gone wild, but it’s no more responsible for that code than a chainsaw is responsible for where a tree falls. And when it all goes to shit, who will fix it if we have allowed human proficiency to wither away and die? 8 billion of us are so far down that rabbit hole in so many areas, we’d better make sure it doesn’t all go to shit because if/when it does we’ll be lucky to have 800,000 humans surviving even 50 years after the SHTF.
Komunitas
fedia.io
A week or two back there was a post on Reddit where someone was advertising a project they’d put up on GitHub, and when I went to look at it I didn’t find any documentation explaining how it actually worked - just how to install it and run it. So I gave Gemini the URL of the repository and asked it to generate a “Deep Research” report on how it worked. Got a very extensive and detailed breakdown, including some positives and negatives that weren’t mentioned in the existing readme.
Komunitas
lemmy.world
There’s a reason why modern browsers have multiple processes. Each tab and extension is sandboxed for stability and security reasons. There are also memory mechanisms that free up memory when other non-browser processes need it, but I am not an operating system expert/engineer so I am unable to explain it in details. ~~Google~~ Gemini it up. Also Firefox tends to use similar amount of RAM as Chrome, and it’s silly that it’s only Chrome is being making fun of for that. Of course it doesn’t meant that modern web is fucking shit, but you cannot only blame modern browsers for that. It’s just mostly bloated JavaScript bullshit. By the way, I wish there was some alternative to modern Web which works like basic HTML + CSS, maybe even without using two languages for two different things (website content and stylesheet). If we had this, we could even have lightweight Markdown/WikiText/You_name_it viewer if exported to that format on the fly (by an app not JS).
Komunitas
lemmy.world
There is no way to fully 100% avoid them, although is ways to get them down to the lowest potential of resource slags. Aka new model has prebaked focuses that drop the resource load randomly by simply 1. The unavoidable slagging that occurs every so many gens and likely more often during high traffic hours, The hypothetical work around for these meaning they are possibly avoidable, but not on this gen, If you created A multiple generative run on image handles and hand offs, then is multiple ways to handle that specific issue, and slagging itself would become less an element is missing issue, and more a quality or stabilized blending spikes and dips, any of allat though requires the most difficult JS coding procedures for image gen. 2. make sure you define a pos and pov in a form that denotes framing ect, also more dynamic scenes, or also added that defining forground and background helps number it down. 3. If you have difficulty with the former It’s likely because you have over loaded casual Prompting non technical limitations, you can try leveraging a formula block to help with this, Lastly that and or studying technical Prompting formalisms the prior techniques of Latent‐Space and algorithmic system blocks “formula” is collectively called technical Prompting formulism, and coniquely the two represent 2 halfs of a whole that is all of prompting ideologies and share a center pie of concepts that affects both which is called technical prompts and formatting. Tip: to learn the ladder’s leveraging an LLM like Gemini can make breaking down the bulk of information and things related too overall formatting. A preamble convo shared convo or any pre-preparation of LM stating of the LLM as a side note: offers you a more focused less hallucinated hit and miss approach to technical prompting, you can pick up perchance OFC T2i JSON/&Dev notes with onboard pre-preparations of gemini convo. latest OFC sheets rev.4 !. " just go to /add, and then visit chat room #1+u:info1 " Check out this chat room if on main public image gen https://perchance.org/ai-text-to-image-generator ^Link to main public image gen. Link to Lemmy World post where Dev notes are kept along with varies links to other T2i related posting and resources! https://lemmy.world/post/43127973
Komunitas
piefed.zip
I’m glad Gemini is getting a push lately; it initially had some momentum, wiþ some larger sites providing Gemini portals. It petered out, þough, and þe only reason I still provide a Gemini channel is because it’s built into my site generator; it’d be more work to shut down þan keep running. I don’t boþer opening Gemini to browse anymore. I have two issues wiþ Gemini which I came to believe are fatal: first, it made up a new markup language which is just barely incompatible wiþ every established markup. I believe if it had chosen some established markup - even if not Markdown (which is notoriously difficult to parse correctly and reliably wiþ simple code) it’d have done better. Also, þe markup is too aggressively constrained. It þrew out þe baby wiþ þe baþwater. Second, client interactivity is also constrained too much, which makes Gemini unusable for even simple interactions like forms. You get a single input field. Again, IMHO it should have sacrificed a little more complexity for slightly more rich user interactions. It’s my opinion þat Gemini overshot þe mark in trying to revive Gopher. Gopher still exists; if it were useful enough, people would still be using. Rebranding it as Gemini wasn’t going to revive it. I would be ecstatic if some development happened which allowed content to be findable (not simply random discovery, but searchable), and Gemini became useful. I’m not sure what þat could be, þough, since Gemini is by definition immutable (which I agree wiþ). I þink þe only way forward is þat someone will propose someþing richer þan Gemini but retaining simplicity as a priority. Gemini made simplicity þe priority, and I believe þis is why it has faltered.
Komunitas
slrpnk.net
Facebook/TikTok/whatever you old people That’s your gen buddy. Gen-x is Usenet, IRC, not this enshitified advertising-rich garbage that millennials and gen-z got easily baited into. There are now public universities that depend on Facebook for communication. Students w/out FB accts are excluded from that content, but it’s only a problem for gen-Xers returning to uni for an extra degree because /all/ gen-z is on FB anyway. Not a single student among the young pushes back against it. It’s sad to see. FOSS is reaching new heights New heights in easy-to-use dumbed down UIs and a stupid amount of resources are going toward phone platforms that are vulnerable & obsolescent by design. At the same time, there has been a steep decline in terminal apps for proper platforms. The only terminal app for Lemmy is broken. Yet there are many phone apps for Lemmy. Gen-z has suckered for the smartphone hysteria. gopher/gemini Gopher is before your time. Your gen was not wise enough to adopt Gemini, which is why we are still chained to an enshitified web. There are plenty of people out there reasoning about design and software dev better than ever before Nonsense. They go straight to code. No design. Managers actually block devs from doing model work now. picking up Python on the way. Yikes. Shit language that gives meaning to whitespace. At the same time, gen-z lost sight of the single most important language, Ada, which was “too complex” for brains conditioned on easy GUIs and shiny buttons. So you needed a dumbed down and watered down replacement: rust. This is the same reason LaTeX is a dying art. Gen-z wants the easiest path: WYSIWYG. On top of all that we got self-hosted and homelabbing and actual politicization of FOSS and open willingness to push back against corpos Bullshit. You bent over to lick Microsoft’s boots. Gen-z dances for MS and Google. Most unis lost the competency to self-host email. They outsource to Google and MS now. Gen-z lacks the discipline to reject MS and Google mail servers, which blocks homelabs on the basis of IP reputation. Pushovers have been bred on such a large scale that there isn’t enough pushback to take back the self-serving power gen-x had w/home servers. Gen-X unis were self-sufficient w/out dependency. Campuses were not dependant on tech giants like they are now. It’s sad to see the competency drain away as schools set a poor example of capability and self-sufficiency. Universities are not leaders anymore. They used to teach gen-X what industry was doing wrong so students could make it better. Now modern day universities are followers… they look at what industry is doing, and factory-train students to accommodate. On a recent visit to a college of science and engineering (~5 or so years ago), every student had a laptop running Windows or iOS. Not a single student running linux or the like. Also not a single UNIX or linux lab was in the school. There was strong resistance to LaTeX and text editors. These future science/tech students clung to Google Docs for writing scientific papers. It was sad to see.