Absolutely. We take for too many things for granted.
Agreed to some extent, but...
Chatgpt is like light years ahead
To put it another way, ...sci-fi.
Was ChatGPT not sci-fi 10 years ago? It's not perfect but come on.
I hope and expect they will be the first step toward a future where we primarily interact with computers using just our voices. Imagine: 'launch Excel, navigate to sheet so-and-so, enter this formula, create a bar chart from such-and-such, format it as we often do'. And all of this in a fraction of the time it takes with a mouse and keyboard.
No idea when that will happen but It would be nice to see a glimpse of it.
Admittedly, haven't watched it, but If AI is dying, why not wait just a little bit longer when it's clearly, irrefutably dead? Would be much more convincing and much less controversial then.
Unless of course, what we're chasing here is controversy.
Did you know Roman public toilets came with brushes that visitors used to scrub their anuses? Yes, they shared the brushes.
I'd say we're REALLY lucky.
Although it's extremely difficult to predict, I think this may be the most probable outcome. However, if we're able to improve reasoning within the next 5 years or so, then we'll have to reconsider most things, starting with the speed of adoption.
Frankly, his arguments for AGI 2024 were not very convincing to start with. And that's me being polite.
Just keep an open mind, don't make big life decisions based on mere predictions, and enjoy the ride. There's lots to look forward to in the coming years.
A sceptic made an absurd comment not long ago that went top, that people in this sub believe we can manifest AGI through sheer power of belief.
So i guess we can try that...
Can't take your comment seriously.
Dude probably read a TechCrunch article
I know! Who does he think he is? One of the inventors of the transformer architecture?
(Yes, he is).
Even if the CEO was one of the authors of the 2017 transformer paper?
Someone with a motive for a crime is not automatically proven a criminal. Likewise someone with an interest to hype AI is not automatically guilty of hyping it. All we may conclude is that being an CEO increases the likelihood of them saying misleading things, but that can be ofset by other considerations such as their background, history, etc.
We shouldn't lazily dismiss claims that don't cohere with our worldviews, then call ourselves sceptic, is all I'm saying.
The conclusion is to be open minded, whether you're a pessimist or optimist. Few things are as certain as we want them to be.
Keeping an open mind is a virtue both optimists and skeptics in this sub could embrace more.
AI Jesus: Sorry but good luck 👍
The optimists should be charitable when interpreting the sceptics' comments but the sceptics should also be charitable and not simply assume that people in this sub "immediately believe everything unconditionaly", no?
'Large Action Model' still makes me chuckle.
If you can be optimistic about the future, you can also be optimistic about your own future.
It will work out somehow.
Imagine then if one of the first things an AGI system might do is to make itself as efficient as the human brain. How many OOMs would it jump by with all that excess energy?
That is François Chollet, though. He's a heavyweight in AI research and sort of LLM critic.
But still, take everything with a grain of salt.
As you can see, the progress is not exponential.
It's over.
I would remind myself it's not a foregone conclusion. I"m somewhat agnostic when it comes to timelines.
And I don't understand why someone would obsess over a fad that they think is going to die soon. Wouldn't it just be a waste of time?
Luddism in the comments reaches peaks of involuntary comedy. 🤯
singularity