ur welcome
312 karma · joined February 28, 2022
https://twitter.com/izzyz
https://twitter.com/ShopifySupport/status/172416611008424349...
If you’re interested, my Twitter handle is in my hn bio.
Theory of mind (ToM), or the ability to impute unobservable mental states to others, is central to human social interactions, communication, empathy, self-consciousness, and morality. We administer classic false-belief tasks, widely used to test ToM in humans, to several language models, without any examples or pre-training.
Our results show that models published before 2022 show virtually no ability to solve ToM tasks. Yet, the January 2022 version of GPT-3 (davinci-002) solved 70% of ToM tasks, a performance comparable with that of seven-year-old children. Moreover, its November 2022 version (davinci-003), solved 93% of ToM tasks, a performance comparable with that of nine-year-old children.
These findings suggest that ToM-like ability (thus far considered to be uniquely human) may have spontaneously emerged as a byproduct of language models' improving language skills.
The steps you elucidated are all expressible in natural language, and we see models like Codex Edit making headway there. One of the most fascinating parts of this is that once access to the known baselines are provided to high-level engineers, they then go on to do much more than what the models alone can do.
The main hinderance to enterprise was compliance but the move toward Azure, etc, will dissolve those barriers this year.
A variation of that argument props up most common AI skepticism. I don’t think there’s anything out right now that would convince you, but from what I know, everything you pointed out will be solved within the next few years.
Codex, AlphaCode, both surpassed by CodeRL on the challenging APPS benchmark last year. Meta working on InCoder. Microsoft working on UniXCoder…
Future research directions are pretty clear from where we stand. That includes iterative methods, reinforcement learning, text diffusion, etc. No one is stuck.
Code generation is moving extremely fast. This tech didn’t freeze in time at Codex or Copilot or ChatGPT. It’s one of the most exciting and difficult domains in AI and the smartest people are all set on solving it.
I’m sorry you’re feeling distress. You’re in good company. A lot of the world is going to have to deal with these problems very soon.
The perfect soldier, citizen, human, is a faceless, egoless pupil.
Most companies probably don't take care of these things at the rate or level you seem to be assuming that they do.
A range input is set up, which returns the number at which the slider or text input is set. This immediately dispatches on any changes.
The diagram cell takes advantage of this reactive dispatching. It passes the result of a chain of ternary clauses, almost like a switch statement.
The cool thing is that you can approach this in a few other ways. You can access an array item if you use the range output as a computed array index, the returned item will correspond with the number set in the range input. Tom added that example as well, in the cell above the diagram cell.
Finally, he mentioned that the user can move the animation at their own pace. That means, using the included Promises lib (not 'Promise', that's a JS module), you can use the .interval or .tick methods to yield a new value at every x msecs. Now instead of having the animation hooked up to the manual range input, you hook it up to the tick method and you've got yourself the basis of a slideshow.
Try it. Just start editing the code, and a copy is automatically made for you to fork. Yes, you can do pull requests on notebooks!
Observable should hire me right?
I'm more incredulous about academically trained domain experts being "brainwashed". Corruption and incompetence are simpler explanations that also make sense.
Breath mindfulness can sharpen your blade, but it won't tell you when or why to use your blade. I present Axelrod as a modern pop culture representation of that.