News for hackers. Stuff that matters.
Re:GPT-4 by xp84 on Tuesday March 14, 2023 @07:16PM (#35160822)
^ Agree. I'm convinced my 2-year-old doesn't operate on a dramatically different strategy than a LLM -- she's learned that when you are negotiating something (continued access to browse pictures on parent's phone, getting to watch TV, staying longer at a place she likes, etc), you can add on "2 minutes?" to your request and sometimes the opposing negotiator will give you some more time. She doesn't know what exactly a minute is or that specific number, but she's observed that it's correlated with getting what you want more than say, a whine. This is simple statistics and probability, in a biological neural network.

I think it's really cute how defensive and dismissive humans get (including those who profess zero supernatural beliefs) when they're trying so valiantly to write off all AI as a cheap parlor trick.

Re:GPT-4 by gerad on Tuesday March 14, 2023 @07:33PM (#35161101)
All that said, the fact that AI is catching up to 2 year olds is pretty impressive. Human's brains surpass dog's at about that age. It shows we're getting close to the realm of "human."
Re:GPT-4 by chromanoid on Tuesday March 14, 2023 @08:14PM (#35161676)
I think finding an analogy with two year olds tells more about those who spout it than about where we are getting close to...
Re:GPT-4 by taneq on Wednesday March 15, 2023 @03:26AM (#35165007)
Given how many university-level tests GPT4 places better than 50th percentile at, I don't know if "catching up to 2 year olds" is a fair description. For that kind of text based task it seems well ahead of the general adult human population.
Re:GPT-4 by Sharlin on Wednesday March 15, 2023 @08:49AM (#35167482)
To be fair, such tests are designed with the human mind in, well, mind, and assume that various hard-to-quantify variables – ones that the tester is actually interested in – correlate with test performance. But LLMs are alien minds with very different correlations. It’s clear, of course, that ChatGPT’s language skills vastly exceed those of an average 2-year-old, and indeed surpass the skills of a considerable fraction of general adult population, but the generality of its intelligence is probably not above a human toddler.
Re:GPT-4 by cornholio on Wednesday March 15, 2023 @02:21PM (#35172501)
You could write a quiz answer bot that is well ahead of the general population without any AI, just by summarizing the first page of Google results for that question. We test humans on these subjects because the information is relevant, not because they are expected to remember and reproduce them better than an electronic database.

If the test is designed to quantify intelligence and is not present in the corpus, ChatGPT does about as good as a dog, and there is little reason to think LLMs will improve drastically here.

Re:GPT-4 by ip26 on Tuesday March 14, 2023 @09:50PM (#35162648)
So if this model has comparable cognitive abilities to your 2 year old, how is it ready to serve as a second opinion for your neurologist?
Re:GPT-4 by mitthrowaway2 on Tuesday March 14, 2023 @10:20PM (#35162891)
It seems likely your neurologist shares a neural architecture with your 2 year old, just benefiting from 30 years of additional training data.
Re:GPT-4 by dinkumthinkum on Tuesday March 14, 2023 @09:54PM (#35162688)
How many watts of power does your 2 year old use?
Re:GPT-4 by flangola7 on Tuesday March 14, 2023 @10:46PM (#35163105)
How many watts does she have access to?

I'm guessing it is fewer than Microsoft.

Re:GPT-4 by astrange on Wednesday March 15, 2023 @03:45AM (#35165116)
That's not the limiting factor since Microsoft isn't interested in paying for you to use the model.
Re:GPT-4 by flangola7 on Wednesday March 15, 2023 @07:40PM (#35176221)
No, I'm pretty sure Microsoft wants you to pay for it, not the other way around.
Re:GPT-4 by melagonster on Tuesday March 14, 2023 @09:56PM (#35162703)
finally we can prove that there are no humanity existing!