No, not even close. We are not "tabula rasa"[1] or blank slate*. If you would actually like to understand why, some good books about this are "The Self Assembling Brain"[2] and "The Archaeology of Mind"[3].
[*] One of the things that frustrates me the most in the discourse on LLMs is that people who should know better deliberately mislead others into believing that there is something similar to "intelligence" going on with LLMs -- because they are heavily financially incentivized to do so. Comparisons with humans are categorical errors in everything but metaphor. They call them "neural networks" instead of "systems of nonlinear equations", because "neural network" sounds way sexier than vectorized y=f(mx+b).
MacOS excels at noise cancellation. Typing on my Windows PC sounds like an earthquake, but you could actually hammer nails during a video call on a Mac and it wouldn't be picked up (it also blocks your voice from coming through, but that’s temporary)
> the people from East Asia, which consists of China, Japan, Mongolia, North Korea, South Korea, and Taiwan.
Seriously? At best, you can maybe if you squint consider Mongols to be white. But the vast majority of the population in East Asia alone is going to be coming from China so it’s a moot point.
For all intents and purposes, East Asia is considered to be people of color.
I wasn't suggesting the like count would be hidden, but the ability to click into the like count underneath the post to see the list of users that liked the tweet. The article isn't entirely clear on if this feature would be disabled once user likes get privated, but I don't see why they would make the effort to do this in the first place if one could still see the users that liked an 'edgy' tweet anyway.
Good! DNNs unlock semantics (parsing, transforming, producing). That's the basis of general intelligence, not encyclopedic random string recall. Models shouldn't burn ungodly quantities of compute emulating DDR5 with their working memory. We need machines that think better, not memorize well. We already have plenty of those.
Massive context windows, and their needle tests, are misguided. We won't reach human-level AGI by basically inventing a natural language RDBMS. Our resources should primarily target better reasoning systems for our models, reinforcement learning, etc.
If we can build a GPT4-level problem solving system that coincidentally also can't remember telephone numbers, I'll consider it major progress.
Memorization usually refers to training data. It's often useful to have something that can utilize instructions losslessly, which is the distinction between these models.
What if your field of vision was infinite and you are looking at a unrolled telephone book?
Would you need a device to remember the phone number? You wouldn't. You would need a method or algorithm to find the number, but there is no reason why that algorithm couldn't be part of the attention mechanism. The attention mechanism is akin to reading the entire phone book for every word you are about to say. It would be unreasonable to expect you to not find the right phone number eventually.
Looking forward to this. My guess though will be that only methods that require a "subscription" will see the light of day. As in, things that require repeated applications, life long medication etc.
Unfortunately, repeated injections cause scar tissue build up. This is the main reason I've steered clear of any injection based therapy. They work fine at first but over time accelerate degeneration.
Doesn't matter. This is leading down a path where if you run a service like Google Translate, you will be expected to restrict certain users (or all users) from translating a publicly-accessible technical report from a security research journal. I can see the same national security logic saying that we can't let security papers be translated to Mandarin, Russian, Korean, or Farsi, because enemies of the United States may use it.
That would be a very oversimplified conclusion. They are categorized as a major defense partner but are also rather close with Russia (getting 65% of their weapons from them [0]). They cooperate strategically but are rather independent. None of this is a slight to India - I'm just contrasting it to Europe which is decidedly in the US camp.
