Hacker News new | past | comments | ask | show | jobs | submit | sfblah's comments login

So you're saying it's something like an updated version of Yahoo Small Business?

My opinion on this is that the damage has already been done. Successive administrations have already loaded up debt massively and participated in a Ponzi scheme with the Fed to pretend asset valuations in the US make sense. Sooner or later the chickens will come home to roost, and valuation multiples in the US will come to resemble those that have existed throughout history, and those which exist in all other parts of the world currently.

When this happens, there will be a big recession, and that is exactly the moment companies will start deploying AI in earnest to save money. This will create a downward spiral.

Yes, the Trump administration is incompetent. But, if you insist on blaming an administration, I would propose Obama and Biden as more reasonable culprits. It was Obama who presided over the massive Fed intervention into markets that enabled unbridled government spending increases. It was Biden who presided over the most problematic parts of the Covid response. Trump is just an idiot. Obama and Biden are actually bad guys.


> that is exactly the moment companies will start deploying AI in earnest to save money.

Companies are already deploying AI everywhere they possibly can as quickly as they possibly can. They aren't waiting for anything. AI just isn't up to the task of replacing most workers just yet. It's still unclear if it ever will be, but they're shoving AI into everything they can to see how it holds up, to see what it breaks, and to give it more data to work with so that it can hopefully be improved.


I strongly disagree with this point. History shows that most companies don't deploy cost-reduction technologies until forced to by economic conditions. When there's a recession, deployment of AI will be next level.


History shows that companies will do whatever they think will make them the most money. They aren't going to hold off on replacing workers with AI if they think that they can just because it wouldn't be nice or because it isn't recessiony enough. The very instant they think they can get away with firing you and replacing you with software they will do it.


Oh? Then how do you explain massively increased adoption of such technologies right after the financial crisis? You're assuming the world works the way people say it does in a business textbook. The real world is messy, and people don't like replacing employees unless they have to


When the economy slows down companies start getting rid of workers regardless of any technology. It's one of the first levers they pull.

Some companies looking to get rid of even more workers might search for new technologies that they hadn't considered before but which could enable them fire more people, or they might accelerate existing plans to replace workers with technology to get rid of them even sooner.

AI is different though because companies are fully aware of it and they are already moving as fast as they can. They have been salivating at the thought of using AI to fire their workers for years now and they're already making every effort to replace workers anywhere it seems feasible to do so. They are in such a hurry that it sometimes results in embarrassing failures like this one: https://www.motorbiscuit.com/airline-ai-chatbots-refund/

Like you say, the world is messy, which means there is no rule saying that companies must leave massive piles of cash sitting on the table until the economy gets bad enough for them to reach out and take it. Shareholders won't accept that. Any company that sleeps on firing their workers when AI can do the job will get absolutely crushed by their competitors who don't sit on their hands.


The Fed intervention was under Bush. The vast majority of the "COVID response" was under Trump. If you look at deficit spending by administration, you will also see that you have things backwards.


Trump 1 increased the debt more than Biden, and Trump 1 wasn't even mitigating a train wreck of COVID inaction left over from his predecessor.


Yeah, I mean people love playing political games with this stuff. But the actual crime was what happened between 2009-2012. Trump is just a symptom. Yes, he's a bad guy, but what do you expect people to do?


I expect people to not vote for obviously bad guys.

Obama too had to clean up a mortgage crisis from his predecessor's careless deregulation. Not an excuse for his own crimes, like say drone striking US citizens. But Trump is just openly corrupt, inept, praising autocrats and aspiring to be one.


Clean up the crisis, yes. Destroy the future of the US economy by endorsing ZIRP and massive government spending for more than a decade? No.

People give Obama a pass when they shouldn't. This stuff didn't start with Trump. The problem goes back much further.


Obama didn't endorse ZIRP. He wanted to do more stimulus but Congress wouldn't let him.

In that environment ZIRP was the only solution. Ben Bernanke (a Bush appointee, re-appointed by Obama) devoted his academic career to how the Fed could have prevented the Great Recession. He was right.


We should hold Obama to account. But at this point I'd settle for enforcing the Constitution.

No need for false equivalence.


Budget deficits are not the same as trade deficits. Trump barely even mentions the debt and clearly does not care about it one bit.


I've been tracking this (literally on paper). I've moved around 75% of my queries to LLMs from search engines. And, the main reason I use search engines is because of queries on mobile, where the devices still make it much easier to search with a traditional search engine; and the omnibar.

Keep in mind that I'm not counting in my 75% queries where I get my answer from Google Gemini I'm just guessing if you added that in, it would rise to 85-90%.

My thought is if browsers and phones started pushing queries over to an LLM, search (and search revenue) would virtually disappear.



I'm a citizen, but I work with folks who are from India and are in the Green Card system. They have application dates going back all the way to 2015. They constantly tell me stories of how the date for being able to get your GC changes and that some people think for Indians it's going to take 20 more years. Can you explain how that works and what's going on there?


Only about 7500 or so Green Cards are issued every year to Indians in each employment based category (capped due to the per-country cap that's applied to India, China, Mexico) in each of the few employment based categories. The number of people who have applied for these categories in the last 10-15 years is much more than 7500 a year, typically almost 10x, so the line moves at a glacial pace, usually moving a month every year.

The line for EB-2 and EB-3 (where most Indians are) is currently servicing those who applied in Summer 2012. So if your colleague got in the line in say Summer 2015, that's 36 months away. At the current pace it could easily be 20-40 years before their turn comes.

Also remember that there was a massive tech boom in the 2014 to 2019 period, so a LOT more people applied during that time so the movement of the line will slow down even beyond the current 1 month per year rate. If your colleagues applied in 2016 or beyond they are unlikely to get it in their lifetime unless the per-country caps are removed.

Finally, you'll sometimes hear the word "retrogression". That refers to the line actually moving "back". This happens because the date that the USCIS announces as the "pointer" to who is being issued cards is an estimate based on the recent green card issue rate.

Sometimes it can move back if they had to issue more green cards than they expected to (since people can apply for spouse and kids together when their turn comes).

Sometimes it moves forward faster if the number of people they expected to apply for the final processing stage turns out lower than their estimation.


This is the most important question. What's the best practice? Firefox? Something else?


If you're looking to move away from Chrome, Firefox + uBlock Origin is still your best bet IMO. Mozilla's committed to keeping the robust ad-blocking capabilities alive despite Google's changes.

Brave is decent too if you want something Chromium-based but more privacy-focused (comes with minor controversies).

Safari works well if you're in the Apple ecosystem.

I actually run a dual-browser setup these days - Firefox for most browsing, and only fire up Brave for those annoying sites that Google has mysteriously "optimized" to run poorly elsewhere. Not ideal but gets the job done!


Librewolf is also pretty good, better privacy than Firefox. If you using firefox, might as well just use that instead.


I’m happy with Zen.


Ungoogled Chromium!


Librewolf


For years I've been reading commentators tell me that QE completely and permanently changed the nature of valuations in US markets. Now, perhaps, we'll finally get to see whether that's actually true or not.

If they're right, no sweat. If they're wrong, a recession will trigger a substantial downward revaluation of assets. For a picture of what that might look like, I suggest reading John Hussman's market commentaries, available free online.


So are you saying that landing the lunar lander in 1969 was substantially harder than SpaceX landing its boosters back on Earth?


In 1969 they used a pilot, and a very good one too.

Surveyor 1 in 1966 is a much better comparison.


Yes. SpaceX absolutely depends on the v^2 automatic deceleration effect the atmosphere provides. They also depend on GPS which is not available on the moon. SpaceX also takes advantage of 50 years of improvement in orientation sensors, INS, image analysis, thrust vectoring servos, and thrust control.

As another post points out, the Apollo landings had real pilots who could look out the window. It's a testament to their skills that they all made it home.


Thanks for the context. Agreed it's amazing what people can do it such an unforgiving environment.


To be clear, it's likely funded by people using Google search as the default on Mozilla.


I thought traceroute always used icmp.


In theory it can use any protocol since TTL exceeded should work with any IP packet. I don't know if there's consistency between traceroute implementations so maybe some use ICMP and some use UDP.


Windows tracert uses icmp.


Depends on whether you mean the tool of the technology. Yes, the common traceroute uses icmp, but since it has become very popular for transit networks to ignore echo requests, there's also tcp (e.g. lft) and udp (e.g. Cisco's default) traceroute tools.


Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: