Everything I described on @ezra klein remains a problem.
👉ChatGPT continues to hallucinate
👉It continues to present untruths with (false) authority
👉It continues to create fake references to support its claims
👉As before, such output can easily fool humans, posing risks
[2/6]
More than that, the thread is based on a confusion, conflating output with process.
A system that regurgitates with paraphrases can sometimes output text that is true, but that doesn’t mean that the system *ever* computes whether what it is saying *is* true.
It doesn’t.
[3/6]
A parrot can repeat something that is true; that doesn’t mean that the parrot knows (or determines, or cares) whether what it is saying is true.
LLMs aren’t literal parrots; but they are terrific at pastiching together *paraphrases* of things they have been trained on.
[4/6]
Pastiches like that are *often* true, because many of the underlying sentences they echo are true.
But they can be false too, because system doesn’t know which pieces belong together. That’s what it means to *NOT* understand.
Tbh @nytimes@kevinroose this seems quite one-sided
👉you scarcely mention the risk for massive amounts of misinformation
👉“commendable steps to avoid the kinds of racist, sexist and offensive outputs” misses eg awful
👉 you are too generous in presuming that long-standing “loopholes” like hallucinations, bias, and bizarre errors will “almost certainly be closed”, when the field has struggled w them for so long.
👉 no apparent effort to ask skeptics (like myself, Bender, @Abebab, etc)
2/3
👉 ridiculous to presume it is “precursor to mass unemployment” when it struggles with math, logic, reliability, physical and psychological reasoning etc.
3/3
*Fabulous* question: How come smart assistants have virtually no ability to converse, despite all the spectacular progress with large language models?
Thread (and substack essay), inspired by a reader question from @___merc___
1. LLMs are inherently unreliable. If Alexa were to make frequent errors, people would stop using it. Amazon would rather you trust Alexa for timers and music than have a system with much broader scope that you stop using.
2. LLMs are unruly beasts; nobody knows how to make them refrain 100% of time from insulting users, giving bad advice, or goading them into bad things. Nobody knows how to solve this.
2. The problem is about to get much, much worse. Knockoffs of GPT-3 are getting cheaper and cheaper, which means that the cost of generating misinformation is going to zero, and the quantity of misinformation is going to rise—probably exponentially.
3. As the pace of machine-generated misinformation picks up, Twitter’s existing effort, Community Notes (aka Birdwatch), which is mostly done manually, by humans, is going to get left in the dust.
4. You are going to need AI. But not (just) the popular stuff. No matter how much data you can collect, don’t count on deep learning. Large language models are without equal at generating misinformation, but they suck at detecting it.
Marcus 2012: “To paraphrase an old parable, [deep learning] is a better ladder; but a better ladder doesn’t necessarily get you to the moon”
@YLeCun, 2022, “Okay, we built this ladder, but we want to go to the moon, and there's no way this ladder is going to get us there”
🧵1/9
Marcus, 2018: “I present ten concerns for deep learning, and suggest that deep learning must be supplemented by other techniques if we are to reach artificial general intelligence.”
LeCun 2022: Today's AI approaches will never lead to true intelligence
Marcus 2018 “I don’t think that we need to abandon deep learning…we need to reconceptualize it: not as a universal solvent, but simply as one tool among many”
LeCun 2022 Deep learning “may be a component of a future intelligent system, but I think it's missing essential pieces”
Just watched Noam Chomsky give a fascinating and up-to-the minute talk on deep learning, science, and the nature of human language.
I loved the first half and find myself deeply skeptical of the second
A 🧵summarizing what he said, and my own take.
Chomsky’s key claims were two:
1. What he wants to understand is *why human language is the way that it is*. In his view, large language models have told us nothing about this scientific question (but are fine for engineering, e.g speech transcription).
I fully agree.
2. In Chomsky’s view, the human system for language is (or comes close to) the simplest, most elegant system imaginable.
I seriously doubt this. Part of Chomsky’s argue for the “strong minimalist thesis” lies in his reading of physics, part of it is empirical.
Let's start with a simple example drawn from my 2001 book The Algebraic Mind, that anyone can try at home: (2/9)
Train a basic multilayer perceptron on the identity function (ie mulltiplying the input times one) on a random subset of 10% of the the even numbers, from 2 to 1024, representing each number as a standard distributed representation of nodes encoding binary digits. (3/9)