Ha ha, thanks, Carver Mead- been awhile since I read his book, Analog VLSI and Neural Systems. He is really brilliant and awesome. Here's his website if anyone is interested: http://carvermead.caltech.edu/index.html
This makes me wonder: China was poised to beat the US in the AI race, why .. why is it that with all their resources and government backing did a llm of chatgpt's magnitude and capability not materialize in China first?
They have true cost advantage in data labelling. So if it comes down to large scale labelling for commercialization they may lead again. As for AlphaZero type of algorithmic breakthroughs, unless Chinese government and corporations are willing to fund that kind of work I don't see why they should lead there.
I used to think Gary Marcus is a voice of reason in the AI field. Now I think he's just a moron shouting hoopla over and over again. Insecurity piling from his work not being relevant any more?
The human in the loop reinforcement learning paper that powered chatgpt's training arose from deepmind's experiments with boardgames (and games). AGI is still an unsolved problem and deep RL that arose from the success of deepmind's experiments wth boardgames and games so far, will likely play a huge part in it
Search? Gmail completing your emails? They have lots of AI products, they just wait until they are more solid before releasing stuff. Now they will rush out a competitor to ChatGPT before it is ready, but they would use it for something given time.
err, I maybe hugely mistaken but is this an example of a real life of the following joke:
"a scientist had this experiment going wherein everytime he rung a bell the grasshopper would jump. He clips one of the grasshopper's legs and rings the bell to see if the grasshopper would still jump and it did. He repeated the clipping process till the grasshopper had no legs. He wondered if it would still jump ... and it did not. He muttered to himself that hence that proves grasshoppers use their legs to hear"
am wondering it this is an example of a teleological thinking?
Yes "random walk" is not a strategy, it is a model. This article basically just says:"Jumping bean motion can be modeled with random walk", which isn't really suprising.
Jumping randomly(edit:in a random direction) when it is too hot is a strategy. And apparently good enough for survival of this species. The bacterium e. Coli uses a similar, or identical strategy of turning("tumbling") in a random direction when the environment conditions are not favorable, and then running straight while the conditions are good. You can guide a blindfolded person to a goal of your choosing just by telling them "hot" or "cold".
Maybe the name of the strategy should be "directed random walk".
Isn't it only a strategy if you've established that they are capable of movement in direct line? Otherwise the most you could say is natural selection selected for shell shape that produces random walks.
I.e., it's a strategy at the species scale, but not the individual scale.
But, maybe that's my mistake of not knowing how these words are used in biology.
It's not the shell that produces random walks, it's the larva of a moth living inside the bean. The larva cannot see the environment, it can sense the temperature, so jumping in a random direction is a means (or a strategy) of the larva with the goal of achieving optimal temperature. It works well in their environment, because the temperature rises if they are exposed to direct sunlight, and a few jumps in a random direction might just get them to shade to a lower temperature. In the shade, they stop jumping.
I think that's right. I don't think the article is imputing "strategic thinking" to the individual larvae. They carry out a strategy, but they probably didn't whiteboard it themselves.
I think it's mainly the ability to start or stop jumping according to temperature. Jumping in a random direction, only when you're too hot, is a strategy for finding shade. Given a pattern of light and shade, that strategy is provably better than never jumping at all, or jumping regardless of temperature. (I haven't proved it, but I think I could.)
I think they're saying a jumping bean "has a strategy" in the same sense that a Roomba "has a strategy." The Roomba has an edge over a traditional vacuum cleaner. But if it's you against the Roomba, you're normally going to win.
"Strategy" makes clear sense for humans. A group of people could think some problem through and adopt a strategy to solve it. For any other life we know of, it quickly stops being meaningful. We may sometimes say that cats or dogs or chimpanzees adopt strategies, because we often like to talk about them as agents. But as we move towards less complex/intelligent forms of life, it quickly becomes clear that we're only ever dealing with biased dice throws.
Point being, I wouldn't worry too much about "having a strategy" vs. "natural selection selecting for". It's only ever the latter - perhaps with exception of humans, but that's up to philosophers to figure out.
Yah, this article is basically nonsense. The beans jump more when hot, and less when cold, meaning they are more likely to stay put when jumping into the shade, and if too hot they'll keep jumping.
"These results suggest that diffusive motion [random walks] in Mexican jumping beans does not optimize for finding shade quickly," the authors concluded. "Rather, Mexican jumping beans use a strategy that minimizes the chances of never finding shade when shade is sparse."
I’m not sure what alternative strategy they are hypothesizing that would find shade more quickly, though?
it seems like the bugs have only one motor function - move in a random direction - and only one input sensor - temperature. What other strategies are available to them than random walk with a probability of movement based on a function of temperature and time?
I don't know either, but is it possible they are concluding that the bugs have only the function "move in a random direction" based on their observations?
If the bugs had the functions 1) move in one direction and 2) turn, then they might have observed a circular or spiral path.
Put a blindfolded human in a zorb on an uneven patch of desert and I would be impressed if they could manage more than one bounce in the same direction in a row. Not sure what mechanism would enable a jumping bean to reliably bounce ‘forwards’.
First, they have no evidence that evolution could create a jumping bean that jumps in a straight line. It seems that the direction the bean moves is a function of the uncurling larvae but also the shape of the bean. Possibly such a movement could have evolved, but the authors are presenting this like the random walk improved the larvae's fitness function.
Second, the conclusion that a straight line would find shade quickest, but only for a small percentage of larvae, is obvious and doesn't result in anything scientifically new. Of course if you draw straight lines from any point in all direction, the straight lines that happened to intersect with the shade will be the fastest path to the shade. But just as obviously, heading in a straight line in a random direction is not the most efficient way to find shade.
If they had the freedom to evolve any search strategy possible, there are plenty of search strategies that will uncover shade faster than a random walk. But the larva don't have the ability to select any search strategy they want, they're blind and have no senses besides their current temperature. They have no way of knowing which direction they're facing. Therefore a random walk is simply all that is available to them, it's not a strategy.
Wrong. The model CAN be characterized as a strategy. Why? Because there are better strategies that can be executed. A model is just a description, a strategy is a method for winning. Some methods are better then others.
Given zero information about your surroundings other then the fact of whether you are currently in a shade or currently not in a shade traveling in a straight line is a mathematically better strategy.
Why? Because random walk includes the possibility of going back to the same place of where there is no shade while traveling in a straight line guarantees every move is a new location.
ad-hoc abstraction is the secret sauce that destroys code bases in the long run and makes things horrendously complex. Composition based on mathematics is more likely to survive years of serving as atomic elements than anything based on ad-hoc human intuition
I forgot where I read this, but I'll paraphrase: it's easier to work with 10 objects with 1000 applicable functions than with 1000 objects with 10 methods each.
In Haskell or other typed fp languages, you have all these well-recognized and well-used typeclasses and their attached methods, conceptually equivalent to generic interfaces in c#/java. It's easier to adapt new data structures into these interfaces and have all your existing functions immediately work with them.
The "mathematics" part here is that these typeclasses are backed by category theory that prescribes their behaviors.
There are plenty of random readings you can find, so I won't recommend them. Someone's bound to argue with me on what's a better entrypoint.
I do want to note that you don't need to use haskell or an fp language to use these concepts. Example: a complete set of fp data structures in typescript https://gcanti.github.io/fp-ts/modules/. It's just that fp languages make using them much easier because of their type system.
> The list, Lisp's native data structure, is largely responsible for such growth of utility. The simple structure and natural applicability of lists are reflected in functions that are amazingly nonidiosyncratic. In Pascal the plethora of declarable data structures induces a specialization within functions that inhibits and penalizes casual cooperation. It is better to have 100 functions operate on one data structure than to have 10 functions operate on 10 data structures. As a result the pyramid must stand unchanged for a millennium; the organism must evolve or perish.
I love programming with typeclasses based on mathematical abstractions. I guess it's because they actually have a chance of standing unchanged for a millennium.
OP there is a lot of potentially dangerous ideas on here. I do not know what your situation is or your mindset, but for the love of all things good and to reiterate what you already probably know. Exercise caution when lapping in these recommendations and stick to basic stuff as much as possible i.e exercise and sleep and mental exercises.
