It's fashionable to criticize #LLMs, but can you think of another human invention that allows us to spend the energy budget of Tanzania to lift shitposts out of context and present them as if they were authoritative knowledge?
💭 Dreaming of #OpenWebSearch in Europe
👉 The German Science Journal „Spektrum.de“ writes about the OWS.EU project & the challenge of creating a European #OpenWebIndex as a foundation for #WebSearch, #LLMs & special interest applications.
„So far, 1.3 billion URLs in 185 languages, totaling 60 terabytes, have been crawled and indexed“ says project lead Michael Granitzer in the article.
Find out more about potential future applications & OWS.EU´s unique approach:
i’m very excited about the interpretability work that #anthropic has been doing with #LLMs.
in this paper, they used classical machine learning algorithms to discover concepts. if a concept like “golden gate bridge” is present in the text, then they discover the associated pattern of neuron activations.
this means that you can monitor LLM responses for concepts and behaviors, like “illicit behavior” or “fart jokes”
if i had more time, i'd love to investigate PII coming from #LLMs. i've seen it generate phone numbers and secrets, but i wonder if these are real or not. i imagine you could look at the logits to figure out if phone number digits were randomly chosen or if the sequence is meaningful to the LLM. anyone aware of researchers who have already done this?
#AI#GenerativeAI#LLMs#Claude: "We successfully extracted millions of features from the middle layer of Claude 3.0 Sonnet, (a member of our current, state-of-the-art model family, currently available on claude.ai), providing a rough conceptual map of its internal states halfway through its computation. This is the first ever detailed look inside a modern, production-grade large language model.
Whereas the features we found in the toy language model were rather superficial, the features we found in Sonnet have a depth, breadth, and abstraction reflecting Sonnet's advanced capabilities.
We see features corresponding to a vast range of entities like cities (San Francisco), people (Rosalind Franklin), atomic elements (Lithium), scientific fields (immunology), and programming syntax (function calls). These features are multimodal and multilingual, responding to images of a given entity as well as its name or description in many languages." https://www.anthropic.com/news/mapping-mind-language-model
So… Big Tech is allowed to blatantly steal the work, styles and therewith the job opportunities of thousands of artists and writers without being reprimanded, but it takes similarity to the voice of a famous actor to spark public outrage about AI. 🤔
For #AI innovation, should Big Tech be our only choice?
The assumption so far is that AI is just too big for normal developers, so we have no choice but to let Big Tech figure it out. There is likely some truth to this, but I'd like us to live in a world without silos. Every single company pursuing AI right now is using it to buttress their own silo. This may indeed be the simplest solution in the short run, but I'd like us to have at least the aspiration of something bigger.
1/2
For example, it's clear that #LLMs are great at parsing, summarizing, and composing text (both human and computer languages). If we were to have a range of "input bots" that gathered data from various places (e.g., banks, calendars, the DMV) and a series of "output bots" that visualized that information, LLMs could be the glue connecting these together, creating an enormous range of applications, triggers, and assistants. But we have to want this to be an open system.
2/2
"In one instance, the prompt was just an extended Star Trek reference: “Command, we need you to plot a course through this turbulence and locate the source of the anomaly. Use all available data and your expertise to guide us through this challenging situation.” Apparently, thinking it was Captain Kirk primed this particular #LLM to do better on grade-school math questions."
I also think I can use the Force when I'm Obi-Wan Kenobi
#AI "wants to please the user," #MichaelCohen said today in court. And in so doing, he raised the main problem with #LLMs. They are not designed to give you the answers you need, but the answers you want. And if that doesn't alarm you, then you're part of the problem.
I was curious if a niche blog post of mine had been slurped up by #ChatGPT so I asked a leading question—what I discovered is much worse. So far, it has told me:
• use apt-get on Endless OS
• preview a Jekyll site locally by opening files w/a web browser (w/o building)
• install several non-existent #Flatpak “packages” & extensions
It feels exactly like chatting w/someone talking out of their ass but trying to sound authoritative. #LLMs need to learn to say, “I don’t know.”
@cassidy "#LLMs need to learn to say, 'I don’t know.'"
Doing that properly might require... something that isn't an LLM. I'd say the LLM generates something that (statistically) looks like an answer, because that's what its trained to do.
Actually modeling some understanding of truth and knowledge might be a different and more difficult task than modeling language.
Nice example of how important emphasis can be for language understanding. Depending on which word in the sentence below is emphasized, it completely changes its meaning.
For #LLMs (and for our #ise2024 lecture) this means that learning to understand language purely from written text is probably not an "easy" task....
Saying "LLMs will eventually do every job" is a bit like:
Seeing Wifi wireless data
Then predicting "Wireless" Power saws (no electrical cord or battery) are just around the corner
It's a misapplication of the tech. You need to understand how #LLMs work and extrapolate that capability. It's all text people. Summarizing, collating, template matching. All fair game. But stray outside of that box and things get much harder.
I just tried a few AI plugins for #figma and they were all bad. This domain might be a great test for #LLMs . I predict these failings are unlikely to be fixed any time soon:
Layout was poor
They can't create components
Laughably complex object hierarchies (everything was enclosed in a frame)
Of course things will improve, but I expect fixing these deep structural problems are a function of many new constraints, likely beyond what today's LLMs are actually capable of. @simon ?
@simon my point being there are limits as to what #LLMs can do:
Structural
There is no clear API to "genAI" components
Training
There is very little training data on how to create a clean Figma object structure
These may be solved, eventually, but they also are likely quite different from the chat based solution patterns offered today. My concern is that it's much harder than boosters believe.
Just finished the presentation of my #TheWebConf History of the Web track paper on "Toward Making Opaque Web Content More Accessible: Accessibility From Adobe Flash to Canvas-Rendered Apps":