>generate generic text responses

Your understanding of the tools is clearly limited so I wouldn't be so pretentious.

Those things aren't mutually exclusive. I can only presume you understand neither.

At this point, it's simply naive to claim that LLMs are useless tools. If you still can't figure out how to use them effectively, that's on you.

That's a bit of a stretch. Top conservationists would be more accurate.

Not even one person asked an energy related question (honorable mention for how many humans can Earth support).

If only Bezos would do something useful instead of this time-wasting, money-wasting endeavor. Jeff, just do something that will benefit society. You've reaped more money than anyone, just give back a little. Build a train. Build a road. Do SOMETHING useful.

Amazon is a free rider. They rely on publicly funded infrastructure to support their entire business, and it's falling into disrepair. The American people are supposed to fund the rebuilding of the roads, bridges, etc. that Amazon needs to survive, but does any single company do more to grind the infrastructure into dust?

This guy should be focused on rebuilding the things his company needs to survive, not shooting rockets into space. All of his projects are myopically driven by his own ego, but what is this guy doing to build anything that actually helps us as a society? He's better positioned than almost anyone to do it, but he wants to build rockets instead.

Consider the differences between modern day tech titans, and those of industry's past. Where's this guy's big infrastructure projects? Where is his railroad? Without us to build everything for him, Amazon would have nothing.

It's time to pay it forward, Jeff.

We're slowly learning that the best use cases for AI involve a human-in-the-loop.

Typically medical experiments compare against the standard of care. In this case, that's a doctor diagnosing a patient from an x-ray. The article is very light on the details of the experiment, but I imagine three groups: 1) doctor alone, 2) model alone, 3) doctor/AI tandem. Medical imaging interpretation is just one area where models perform extremely well (the training data is well suited to the problem). It's irresponsible to completely outsource the diagnosis to a model, but a human-in-the-loop system is clearly time-saving, cost-saving, and apparently patient saving.

I work in a completely different field, but we are already using models in analogous scenarios, and they improve prediction compared to human-diagnoses alone. This isn't a surprising result.

Writing headlines should at least *try* to capture the gist of the article.

"Humans Beat GPT-4 in Turing Test by Small Margin"

The entire site exists to suckle at the teet of tech industry titans. Every other post is in some emoji clickbait format, seeking to glorify the worst of corporate America.

... for what is considered "intelligence." A lot of comments about "this is just pattern recognition" seem to miss the point that most of human cognition is pattern recognition.

In fact, I bet most of you poo-pooing these comments by Masayoshi Son couldn't even give a proper definition for intelligence (without researching a specific counter example) that deviates substantially from what GPT-4 is already doing. And in that research process, you would probably find that GPT-4 can provide the same -or better - answer to that question than you, in less time.

I recommend Bostrom's book "Superintelligence." We're not that far away from AGI, but we might be quite far away from ASI (artificial super intelligence) which many of you wrongly consider to be the only kind of AI.

The morality of a website that awards dollars for fulfilling someone else's confirmation bias is truly something to consider. A glorious echo chamber.

That's the point. It keeps you in the ecosystem for a few more minutes just by hovering and glancing around.

From a genetic point of view, there's no short term selective pressure that would reduce the incidence rate of psychopathy and sociopath. Psychopaths and sociopaths are are still out there, but maybe they are finding other release valves for their urges. Like politics.

> "Just imagine no commute to the office in the morning and instead, you could have an extra hour of sleep and less friction," the description reads. "Next, you could walk out of your room and quickly grab a delicious breakfast or get a workout in before work starts."

It's just so patronizing.Yeah jackasses, you're making me come into the office, probably for no real reason other than self-satisfying, company-wide metrics. And yet you still can't help yourselves in writing corporate vomit like this.

Of course you can't really give up your lease either right? So ostensibly, google just wants you to pay more for the same convenience you would have just by working from home. You can't advertise a service the employee pays you for as a perk. A cheap hotel room so you don't have to commute? It really shows how little google is willing to do for people. We're going to do nothing, you're going to pay us, and you're still going to come into the office.

The absolute number of people killed is a really poor metric. You should be more interested in accidents/deaths/property damage amounts per mile driven or per thousand miles driven.