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

White collar jobs will be destroyed by AI/AGI. But the US workforce consists of a large portion of white collar workers.

To replace the coming destruction of white collar jobs, manufacturing and industry jobs must be brought back, along with new resource collection initiatives (e.g. mining). They will be needed anyhow to build robots, drones, and other machines to compete with China and India, technologically and militarily.

Globalization will be effectively dead, remaining trust in the petrodollar will be destroyed, therefore so will Pax Americana. Expect more wars.


There have been rapid recent advances in robotics, both classical industrial robotics (advanced by rapid iteration with digital twins and simulation environments like Omniverse), and humanoid robotics (pushed forwards by Boston Dynamics, Figure and a lot of other recent entrants). So if we really do achieve AGI that would take over white collar jobs, manufacturing jobs will likely too be taken over very soon thereafter.


In your scenario white collar jobs will be destroyed globally.

So effectively you would be paying people to work in manufacturing even though it's no longer necessary.

You may as well pay a basic income instead.


The article is titled Why I don't discuss politics with friends but it doesn't explain the why? Unless I missed it. It seems to just talk about the challenges.

Why don't you discuss politics with friends? Are you worried about loss of friends? Do the conversations ruin your day? Do you feel alienated?

Depending on the why, there's different points I'd argue for or against the reasoning. Without that piece, it's kind of hard to discuss the premise of the article without just guessing its implications.


This sentence was intended as that answer, but I guess it wasn't clear enough:

"And this is fundamentally why I don't discuss politics with friends.

It's not that I don't want or am scared of opposing views (in fact the opposite is true[8]), but rather because of how common others’ desire to "remain in the bubble" is."

I actually am willing to risk alienation to find people that enjoy this sort of discussion-based discovery as much as I do, but found most people I encounter don't actually want that -- so I try and respect what seems to be the average opinion.


You might want to add these HTTP headers to help reduce the server load from your complex web application:

Cache-Control: public, max-age=2147483647, immutable

Expires: Wed, 31 Dec 9999 23:59:59 GMT


I just flagged it since it's politics. Don't need to get much further than seeing the comments here to understand why posts like these don't belong on HN.


Use your hide function, this is important.


Hide doesn't flag content that should be flagged


So you want to hide it for everyone.


Flagging is for flagging content that doesn't meet HN's guidelines:

https://news.ycombinator.com/newsfaq.html#flag

Specifically:

> Off-Topic: Most stories about politics, or crime, or sports, or celebrities, unless they're evidence of some interesting new phenomenon.


> unless they're evidence of some interesting new phenomenon.

Like the death of American democracy, the death throes of the global status quo? The normalisation of facism?

I've been here for a decade and a half and I've never once felt the need to discuss politics here before. This situation is very different and is sure to have a direct effect on anything else we may want to discuss in the future. There's also often a technological aspect where I am dying to hear expert opinion and this is where the experts are.

Again, this situation is very new and it requires a response. Hundreds of people are upvoting these articles every day and every day they're being told to go elsewhere. It's a very disconcerting feeling and I don't understand why "hide" isn't enough.


Dang and other moderators can decide whether this submission qualifies as an "interesting new phenomenon" and turn off the flags like they sometimes do. Given how frequently these political posts have been showing up on HN over the past few months, I'd wager this one is not an "interesting new phenomenon" to them.


This is more of a solved problem than not these days thanks to LLMs. You can plop an agreement into an LLM chat and ask some questions, which is a lot better than just checking a box because you didn’t have time to read it. I’ve been doing this myself regularly with pretty good results finding things to be concerned about, or not. LLMs hallucinate and aren’t equipped to be attorneys for us, but this is a big improvement over just having to accept everything blindly.


The predecessors of agentic AI, humans, have figured this out. They use this thing called user interfaces to do computing. If agentic AI can't use UIs like humans, why bother calling them agents?

I feel like there is momentum around this idea that all of computing needs to be re-invented (e.g. MCP, "agentic AI needs its TCP/IP moment") to address the deficiencies of agentic AI when what really should be addressed is their inability to be true agents for humans, using the tools humans already use.

Using UIs doesn't even have to be visual. Visually impaired users regularly use computers as well. Since UI for the visually impaired is all text based, it seems reasonable for LLMs to start there.


Didn’t Firefox rise from the ashes during the years of Internet Explorer’s waning monopoly? How did Firefox “take back the web” pre- Google funding?


It didn't. Mozilla were part of AOL until 2003. Google funding started in 2004. They released Firefox 1.0 in late 2004. And web browsers were much simpler then.


I think Google funding was there from early on. (Though I suppose it’s also because there was less corporate bullshit going on in Mozilla’s early days.)


It was a lot easier to build and maintain a web browser back then than it is today.


Could be a good opportunity for using an LLM to summarize and extract anything important?


Here’s a critical NOTAM that was missed. Translated, it means, if you fly here, the Russians will shoot your ass down. And so they did. But nothing encodes that information in the NOTAM so there is nothing for an LLM to summarize and extract. Expecting an AI to compensate for poor system design is magical thinking.

A1492/14 NOTAM Q) UKDV/QRTCA/IV/BO /W /260/320/4822N03807E095 A) UKDV B) 1407141800 C) 1408142359EST E) TEMPO RESTRICTED AREA INSTALLED WITHIN FIR DNIPROPETROVSK BOUNDED BY COORDINATES : 495355N 0380155E 485213N 0372209E 480122N 0370253E 471352N 0365856E 465018N 0374325E 465900N 0382000E 470642N 0381324E THEN ALONG STATE BOUNDARY UNTIL POINT 495355N 0380155E. RESTRICTION NOT APPLIED FOR FLIGHTS OF STATE ACFT OF UKRAINE. F) FL260 G) FL320)

Besides that, would the developer of the LLM accept liability for accidentally filtering out important NOTAMS, or hallucinating NOTAMS that did not exist?


For those who can't read NOTAMs: This is a NOTAM that was in effect during the MH17 shootdown, closing the airspace from FL260 (26,000 feet) to FL320 (32,000 feet.) In the previous three months, 16 aircraft were shot down in Ukrainian airspace, including one three days before at FL210, which is what prompted the NOTAM. Ukrainian officials should have closed the airspace entirely, but chose to keep it open and issue this NOTAM instead. MH17 was at FL330, so in fact it did not violate this NOTAM at all.

And for some further context: the quote upthread of "NOTAMS are a bunch of garbage that no one pays any attention to" from NTSB chairman Robert Sumwalt came after Air Canada flight 759, which would have had >1000 fatalities if two airplanes were a few feet closer: https://en.wikipedia.org/wiki/Air_Canada_Flight_759


I mean, I just copy/pasted that into ChatGPT with no other instructions or context, and it seemed to handle it pretty well:

This NOTAM (Notice to Airmen) establishes a temporary restricted area within the Dnipropetrovsk FIR (Flight Information Region) in Ukraine. The restricted airspace is defined by specific geographic coordinates, forming a boundary for the restricted area.

Key details include:

- *Effective Dates:* From July 14, 2014, at 18:00 until August 14, 2014, at 23:59, with the "EST" indicating that the end date might be subject to change.

- *Altitude Limits:* The restricted area applies to flights between FL260 (Flight Level 260, approximately 26,000 feet) and FL320 (approximately 32,000 feet).

- *Exemptions:* The restriction does not apply to flights of state aircraft of Ukraine, meaning those aircraft can operate within the restricted area during the specified timeframe.

Pilots and airline operators should be aware that flights not exempted by this NOTAM must avoid this area or adjust flight levels accordingly to comply with the restrictions. For further guidance, they should consult official aviation authorities or current NOTAM listings, considering the potential for changes.


FWIW, this exact text was probably part of ChatGPT's training data, but the next important NOTAM might not be.


Perhaps. It would be easy to past in an arbitrary one and see how it does. I'm pretty sure it has information on how to decode these in general in it's training set.


That would be impressive because there isn't a standard format for NOTAMs. They are intended for humans, not machines. It's a major problem that safety critical information is often not available in machine readable format, so you can't have a plane's navigation system automatically give warnings if you program it to fly into restricted air space or to a navigation beacon or runway that is announced as offline for maintenance by NOTAM.


I have an AI-powered NOTAM mapper at https://notam-mapper.obliscence.com. It's just intended for fun, but you can try pasting arbitrary NOTAMs into it and check the results.


That looks less like poor system design (although I'm sure that's contributing) and more like understatement that's practically criminal. Who the hells hears "Russia may shoot down your plane if you go here" and thinks they should convey it via "tempo restricted area bounded by coordinates"? That's like posting a "do not enter" on a door with a bomb behind it. Is English so hard?


For pilots, "restricted area" is already synonymous with "absolutely do not even think about going here" and is used for things like military testing and training areas that regularly use live ammo, extremely sensitive military or bases, etc. As a concrete example, Area 51 is in a "restricted area".


> For pilots, "restricted area" is already synonymous with "absolutely do not even think about going here"

I don't buy it. How often do you hear of a passenger plane getting shot down for entering restricted airspace over peaceful territory like the US? What about over combat zones, like over Ukraine or Russia? Who in their right mind equates these as as equally severe? If you had to make an emergency landing would you seriously view these equally?


There are more risks than getting shot down. All sorts of dangerous activities may be going on in restricted areas: missile launches, dogfights between fighter jets, tests of unmanned vehicles, etc. You don't want to be a surprise guest at any of them. "Restricted" is a single, short word taught in pilot school to get the message across clearly. If a sudden emergency requires entering a restricted area, then air traffic controllers are there to coordinate it.


At least within the US, entering a restricted area without authorization can result in your own personal military air show, followed by lots of questioning, possibly prison, losing your license or both. You don't mess around with restricted areas unless you have a damn good reason.


>> I don't buy it. How often do you hear of a passenger plane getting shot down

> At least within the US, entering a restricted area without authorization can result in your own personal military air show, followed by lots of questioning, possibly prison, losing your license or both. You don't mess around with restricted areas unless you have a damn good reason.

You're not taking my point.

I don't know about you, but I would sure as hell prefer prison over myself and hundreds of other people falling from 40,000 feet.


> ...I would sure as hell prefer prison over myself and hundreds of other people falling from 40,000 feet.

On the one hand, that would be ideal.

On the other hand, that appears to be a NOTAM restricting access to an active war zone. It is truly tragic, and should be avoided whenever reasonably possible, but sometimes civilians who enter war zones get killed.

So, (just so you know) the "they flew into an active war zone" bit is probably the reason why your point is being brushed aside. Restricted airspace is restricted airspace. Unless you're on the unrestricted craft list, you simply don't enter it.


To my reading it's very understated if that is the entire text of the NOTAM - I would have liked it to say WAR ZONE or GUIDED MISSILES in there somewhere.

Compare it to the DC SFRA[0] which specifically states "MAY USE DEADLY FORCE" even though that's extremely unlikely to happen.

When Flight Restrictions are active in FAA controlled airspace, ATC will often clear you through them if you're on an appropriate flight plan without you needing to ask. Even the 'deadly force' SFRA above is essentially handled for you so long as you file an IFR flight plan.

[0]: https://www.faa.gov/air_traffic/publications/us_restrictions...


When I'm flight planning and I see a restricted area along my route, I honestly don't need to know why. I'm not going there unless the controlling agency gives me authorization and I understand the danger. Because there is danger; otherwise it wouldn't be restricted.

If it makes you feel better, here's the current TFR over eastern Ukraine. It even explicitly says "danger" - does that make you feel better?

TEMPO DANGER AREA ACT WI COORD: 434800N0392400E-433000N0303500E-441500N0302500E-...


Yes, I didn’t know my phone had a secret feature of becoming an instant hand warmer until I visited this website


If the goal is to build AI agents, I’d think the maximum utility from them would come from building agents that can use the UI that humans use. Or at the very least, the lesser available API. MCP is yet another interface to services when there’s already at least two out there available to use


One thing I like about MCP's decision _not_ to just use HTTP APIs is that it makes it possible to build and distribute tools that people can use without having to set up a server for them. This goes a long way for enabling open-source and locally-running 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: