Comment by IIAOPSW | Hacker News Reader

Comment by IIAOPSW | Hacker News Reader

Bold and seemingly contradictory predictions by me:

-The rapid advances in AI slow down. Dalle and GPTchat will be an exception not a rule in the long run. The fraction of hard AI problems that can be solved with raw force of more parameters and training isn't that large.

-Paradoxically, AI will change everything. The advances that are already here haven't caught on yet. Legal documents in particular occupy this sweet niche of having some of the rigidness and strictness of coding but without the exact syntax requirements needed for code. The lawyers haven't yet noticed the machine coming for their job. I'm sure lawyers are not the only ones.

-A great recession isn't coming. The markets just sort of trade sideways for a while. You won't get 9% annual off an index fund like you did in the past, you won't loose your shirt either.

-The big issue on everyone's mind is the war in Ukraine. From what I've seen of just crude projections of when Russia would run out of tanks, APC's, whatever else, Ukraine will win the war outright in the first half of this year.

-China has basically given up on zero covid. I say this as someone who earlier was planning to speculate on China doing another major lockdown: they ain't going back to it.

-First the plague, now the war, next the famine!

100% lawyers have noticed. For example: they've been trying to automate discovery for a while now. I'm sure the new LLMs are making waves in those circles.

But also: lawyering is good money, and they're adept at navigating the law. So breaking into that space with some sort of "tech disruption" has a high chance of getting you (metaphorically) dismembered.

Yes, my firm is investing loads into AI and other tech solutions to common legal problems. I first worked with eDiscovery software eight years ago. More recent trends including using machine learning to (partially) automate legal due diligence.

Liability is the issue, as you say. If the software you use fails to flag an obvious issue that you consequently fail to flag to your client, you are in trouble, and the software provider won't be indemnifying you for it.

One of my predictions is that this will lead to an AI arms race. The front line won't be weird melty pictures and code that doesn't work but looks like it might, but increasingly expensive automated suits by legal AIs against other legal AIs.

This will - ironically - create the Singularity. And then we're all either in jail, bankrupt, or legally executed.

I'd love to talk to you about it if you're willing. Email's on my profile.

Yep. My buddy actually worked at making one practice less paper dependent. If there is a place, where tech did not make a big impact yet ( and some of it for a reason like confidentiality ), this is it. Naturally, doing it well will be hard ( and suddenly finding a lawyer that knows what he/she is doing tech wise will be about as bad as .. any other company on the planet ).

Most of these are pretty good, if not a big stretch.

>First the plague, now the war, next the famine!

I'm noting no predictions on this page regarding the Colorado River and the drought in the US West. Will the drought continue? Get worse? Could it cause a famine? Or just widespread adoption of efficient watering systems?

There's more than a few "oh shit the food supply" stories sizzling in the background. Wheat exports from Ukraine/Russia is an obvious one. Spotted lantern fly eating the crops we like to eat in the American North East is another. Colorado River drought is yet a third. They might all be nothing, but only one of them has to be something.

The west coast was desert once and to desert it will return.

Drought is hitting Europe and Asia as well. Water will be traded as a commodity if it isn’t already beyond water purification companies.

Food prices in North America will continue to climb as drought worsens agriculture in California and Mexico. California supplies most of the food in Western North America, including the Western half of Canada

Yeah it's been hitting even central Europe (Germany, France, Benelux AT/CH etc) for the last few years.

It's harder to notice from afar because it's not deserty (yet) but the temperatures are going up all year and with it the obvious problems and also new and "exciting" fauna and flora is arriving.

Many places in that area are historically very water rich that's why the consequences to everyday life are often not yet that severe.

No idea about Europe but I think the Colorado River watershed will improve in 2023, ruining any chance of getting meaningful change accomplished - and guaranteeing deadpool by 2030.

For the record, the National Weather Service predicts average snowpack in the headwaters of the Colorado.

Nice prediction but I think that lawyers will be the last to feel the change. You don't pay a lawyer for crafting a document, you pay a lawyer to take responsibility for the document and to foresee any possible problems.

Oh you'll still pay a law firm for the responsibility of the document, but the firm is going to employ far fewer people for the same work.

China has already given up on zero covid. Reports are that they had 37 million cases last week.

I believe AI will be one of those things where the first 80% is the easiest and the last 20% very difficult.

We probably need at least another doubling of global processing power before we will see any major advancements in AI tech, which will take 3-7 years in my opinion.

I would enjoy being proven wrong, but I fear I am actually overly optimistic.

The real moment of truth will be if any models start to assist massively in research in the hard sciences.

Based on the quality of outputs I get when asking for help with somewhat complex AI research problems, I think it'll likely help accelerate the pace of other research as well, and discovery will be limited by people's speed of running the tests it suggests and feeding it back the results.

For 2023, AI will go nowhere, even if developed further. It will remain a gimmick. Expectations are too high, and AI research can't deliver.

+1 for for the recession, it's not coming, that almost apparent right now. Yes, more will lose their jobs, but only because many of those positions should never have been created in the first place. There are still plenty of jobs, just less attractive ones. I will say: I have zero feel for the US job market, it's simply too weird to comprehend. For Northern Europe, plenty of work all around.

Russia won't run out of APC, but they will run out of tanks, they basically have already. They don't have the inventory to deploy tanks in offensive actions anymore. They won't run out of APCs. The will run out of good one and competent personal to put in them. The last more of a statement of current facts, not a prediction.

I have a friend who has been in china since before covid hit, teaching english to kids.

He just told me today that China has lifted all restrictions surrounding covid and are letting it run its course. As a result, he and his gf have both gotten covid for the first time.

When it first hit they had tanks driving up and down their streets making sure people stayed in their houses, so it's definitely a shift for them.

Reply on news.ycombinator.com