Some good points, but unnecessarily tainted by comedy punchlines.
I want more people bringing up discussion points without offering a grand solution or taking a side. I have no bloody idea if AI is a good or bad thing in the long run. Or if it's "fair" that it takes away artist jobs by using their previous work. Or if it will take away jobs - and if so, if that'll end up being a good or bad thing. It's still a good discussion to be had. Predictions about the future have a notoriously bad accuracy if you look back
my takeaway is that our tech overlords praise science to the sky while making statements that are not scientific at all and bet their enormous resources on it.
I respect a lot of political critique of Job Stewart and his ability to communicate to the masses in simpler ways. But, I felt this was way off the mark. It loses the logic in order to bring out the comedic appeal.
Sure, there are downsides to a new technology that improves productivity. But, it is a tool after all. What are we supposed to do? Not research anything new because it might put somebody off their job? Keep mining coal because the coal miner might lose their job? Keep running ICE because the workers at an ICE factory might lose their job?
There is always a cost to progress. It is better to talk about managing the progress (regulate that all AI needs to be open to analysis, publish their weights etc, tackle reproducibility of results etc.) instead of dumb fear mongering. Even Job Stewart cannot stop LLMs from being improved.
I didnt take Stewart's monologue as anti-progress. Rather the core takeaway is the vague and grandiose talk of greatness and utopia for all of humanity and the planet that is to come from the AI. And brushing aside the upheaval and consequences mass automation on a decade scale, not multiple decades, will cause for the N% of society affected negatively. Big Tech personal AI assistants aren't going to put food on people's table after their profession is automated away. And political nonsense already make it hard for many to get aid when they're unable to work as it is, imagine some politician saying go get a job, don't leach off the government and tax payers.
Is there a way to minimize the negatives that will inevitably playout? Those AI leaders aren't talking a lot about that, only vague suggestions or passing mentions, and then back to "we're going to use AI to make the world a better place"
> Big Tech personal AI assistants aren't going to put food on people's table after their profession is automated away.
The future will be great! Everyone will have a supercomputer in their pocket, loaded with an AI therapist program to help treat their depression resulting from being forced into poverty and having to live in a cardboard box under a bridge.
Sorry, but that is not fair or even normal. Are you saying that anybody can do something new only if they _can_ handle all the externalities that they cannot possibly have control over?
Lets say you develop a new, better calendar tool that makes secretaries/personal assistants redundant. Are you supposed to release it and make money off it only if you could find a way to employ all those personal secreataries and assistants in the world?
I do agree with you and hate the trope that is "we make the world a better place" or using the word - sustainable and green color to things that cannot possibly be either.
If a politician says "go get a job, don't leach off the government and tax payers" and is so out of context, may be the political system and electing such a politician is a problem, and you ought to improve the society than asking Microsoft to think about the jobless...
Yes, if you create something incredibly dangerous to current society, you should not be misleading the public about the negative externalities. Oil companies who knew that pollution caused global warming and buried it should have every single C-suite and boardmember put against the wall and shot dead.
> Are you saying that anybody can do something new only if they _can_ handle all the externalities that they cannot possibly have control over?
No, not at all. But to focus majority of the time focusing on greatness and downplaying the consequences will skew the public perception. If you're skewing perception then you, as a society, aren't fully understanding the implications of what's happening (eg social media in 2010 vs social media in 2020s, mental health, people quiting, etc).
> If a politician says "go get a job, don't leach off the government and tax payers" and is so out of context, may be the political system and electing such a politician is a problem, and you ought to improve the society than asking Microsoft to think about the jobless...
Yup, there's definitely a issue with current political systems. The system is likely going to resist change and I'm not foolish enough to think it will change much any time soon given the current trajectory. Obvious this varies country to country.
> I do agree with you and hate the trope that is "we make the world a better place" or using the word - sustainable and green color to things that cannot possibly be either.
That's all I was highlighting. It's all grandiose promises of sunshine and rainbows as one may naively assume if you take any AI "thought leader" at their literal word. I'm not going to quote any of them, but the clip this post links to has a number of direct quotes from them about what great planet scale problems AI will solve.
I wasn't proposing any solution, but trying to highlight the naivety of assuming AI will usher in an era of greatness for _all_.
It would be great if you could get all the benefits and humanity could live on this planet, and maybe others, for millions of years in prosperity and greatness, but the systemic underlying issues will continue to persist even into this next era.
I remember watching one of his show's episode on Apple TV about GameStop and somehow he thought the problem was the (deliberate) obscurity of modern financial systems (instead of, say, how stock trading creates wrong incentives for a healthy economy). That was pretty weird.
You set up an economic system that ensures that all of humanity benefits from the progress of technology and not just capitalists like how worker wages have stagnated in the last 50 years while productivity has grown incredibly with the internet and computers.
Some basic social systems are ensuring Basic Income is available so those displaced and cannot find jobs aren't in hopeless situations.
The government needs to alleviate the tax burden on workers and push it more (eventually totally) on corporations as they become where all profit is made. This means ending loopholes and threatening serious jail time (5 years for minor; 15 for major offenses) for any and all C-suite and board members who try to abuse the system. Fines are just the cost of business. These people need to be fucking terrified that they will be raped in jail for the rest of their shitty old lives. Whistleblowers should not be getting fucking killed in this day and age - its a fucking embarrassment and every politician who could have done something and failed should be put against the wall and shot dead.
Lobbying has to end. Any politicians, lobbyists, C-suite/board members who engage in corruption should go be terrified of going to jail. If you decide to be a public servant then you are now under as strict watch as police should be. All your communications are monitored and you have to wear a camera everywhere you go. And the organizations who watch have a smart (smarter than I can come up with) series of checks and balances to ensure they cannot be easily corrupted either.
Government in general should be redesigned to limit certain capabilities and focus on what is key like regulation of AI. We need scientists being the focus as the voice of reason. Not putting shitty, corrupt politicians in charge here.
As a former Dr of Mopping, how dare Jon elevate that profession so highly. It's truly a shitty job.
But he has a point, cars robbed everyone of transport by carriage. And maybe one day ai bikes will rob cars of their use. Despite the doom and gloom occasionally I have faith in humanity.
Apple asked Stewart not to have Khan on a podcast.
Nor would they let him do the comedic segment on AI.