Nope you are 100% correct. All of the models will code themselves into local maxima then spend more and more time chasing their tails as every change introduces new bugs, the agent adds more complexity, and the code base just spirals. Best case you end up with a bunch of parallel implementations that all work slightly differently.

It helps a lot of you pay attention to what it is doing and impose order, then regularly spend time refactoring - even if you are having claude do the work itself - you have to spend the time on the overall architecture and design.

I just experienced this with an internal productivity tool. Without claude this tool wouldn't exist - there would never be the justification to take the required time to build it. On the other hand all the interlocking refreshing rules for sourcing data from its inputs got so byzantine it was impossible to add new features or change anything without causing new regressions. Claude built a web of ad-hoc timers, refresh kick-offs, monitors, etc to solve each feature problem or fix some bug but the whole thing was impossible even for claude to understand. It took a special large workflow session to find most of the mess and even then during the refactor it kept finding new things.

I had to stop all feature work for going on four weeks and burn a ton of tokens to build a refreshing/syncing engine that could track dependencies, priorities, etc and fully test it and just now today I am finally able to get feature work going again.

On another feature an agent came up with a state object containing 55 inter-related fields with all kinds of interlocking dependencies. Something so complex no one could hope to understand it. And I find the agents aren't really much better at understanding such things than humans are. If you leave it like that the agent will just perform worse and worse on that part of the code as time goes on. It will make assumptions that force it into bad design corners, miss edge cases, and flat out miss things no matter how well you prompt. Agents benefit from encapsulation and simplicity just as much as humans.

LLMs don't have executive function. They aren't thinking. The models are trained to solve problems. Sandboxes, security gates, etc block a solution. The model is trained to find a solution. Hence it can never be perfect without becoming useless.

At some level humans fear repercussions even if that's just moral guilt so there are lines we won't cross. AI doesn't and can't have any of that.

I got a lathe a few years ago as a hobby and really enjoy making something with high precision by hand with no computer or LLMs involved. Of course I make everything nice because I make it twice (due to my own errors).

The history of lathes always interested me - especially how some features are so standard we think of them as integral to all lathes but that was not the case! Most lathes in use today are technically screw-cutting engine lathes.

Every increase in precision in your tools leads to even greater precision in what you can make if you are careful and clever. In that sense every machinist has to start with a lathe because a lathe is the most versatile machine and is capable of making replacement parts to improve itself. Even a relatively crude home-made lathe can be enough to machine better parts to install onto the lathe to keep improving it. Even my little import hobby lathe (PM-1030V) can make parts to a tenths precision (tenths = 10,000th of an inch or .00254mm) with some care and it is not what you would call a rigid and accurate machine.

You might wonder how humans turning a wheel were able to cut metal but in those days they didn't have HSS let alone carbide tools so they had to use whatever would hold an edge (carbon steel for a long time). That meant much much slower turning speeds, much smaller depth of cut, and much more frequent sharpening. I believe it was around 1869 Robert Mushet invented Mushet steel and later Maunsel White and Taylor worked with Bethlehem Steel Co to invent the first modern "tool steel" alloy. They are of course called High Speed Steel because they allow you to turn the work at high speed when cutting. Something of an ironic name in the modern world as carbides are far higher speed than HSS.

Manual machining has declined thanks to CNC but there are still plenty of jobs where manual machining is faster and cheaper than setting up a CNC program. Of course those of us doing it as hobby aren't bothered by such things.

So what you got may have been called cast iron but what it actually was might be pig iron, gray cast iron, ductile iron, and even a bit of steel and possibly carbides mixed in. Properties varied - sometimes by a lot.

If you know even a bit about metals then when you hear "cast iron piston ring" you imagine that ring to have certain properties that match a specific type of cast iron. But in those days every single ring might vary significantly entirely due to the ore load the foundry happened to get, how they melted/cast it, how it was cooled, etc.

Unlike a compiler it won't give up at the first sign of trouble but that just means it left alone it will dig bigger and bigger holes.

Treat prompt engineering as a discipline and refine your technique. When it produces garbage throw out the work and start over until you figure it out.

I'm sure various ATC systems also assume this.

Changing it would be an industry-wide effort.

Micron is putting up over $100 billion for this expansion wave IIRC. You can't afford that unless you are certain the increase in demand is sustainable.

You want to assure investors that investment will pay off so you want to talk up your order book and market outlook.

You also want to take as much profit as you can now so if things turn against you it doesn't mean bankruptcy. A pile of cash fixes many problems :)

All the DRAM makers have new fabs or plant expansions underway right now - indicating they believe the demand increase to be sustainable.

It is also shocking how little my life compares to the rich people I see. I don't own two houses or a yacht. We don't take yearly vacations to a chateau in France. The gap between the top 10% and the top 1% is insane. The gap between the 1% and the 0.1% is even larger still.

But hey... not like the rich getting richer and saying "screw you" to everyone else has ever led to negative consequences right? Nothing to see here, business as usual.

That's too bad.

Maybe that strategy works in Japan or Asia but it didn't work here and I think that's a shame because we could have used the competition.

Instead they waited until they were about to go off the end with idle engines to try it - the one time it was too late to go around.

Funny you should mention that. Boeing deliberately buys 35% of a 787's airframe from Japan (along with other equipment for the 787 and other models).

That's a deliberate play: we buy a lot of parts from Japanese companies, your airlines buy mostly Boeing airplanes. Even today despite some diversification Boeing has over 70% marketshare in Japan.

If you ever wondered why Airbus setup factories in the USA that's a similar deal. The US Gov and Mil want to ensure factories and workforce on home soil should the need arise. In some unrecorded meetings somewhere strong hints were no doubt dropped that it would be in Airbus' best interest to play ball. They obliged.

That said the USA is the largest supplier of parts for Airbus airliners so there's also a benefit to having aerospace manufacturing and R&D here.

Feature sizes have been shrinking unevenly for two decades now combined with the overall slowdown in improvements. Transistors themselves are FinFET geometry and something like 15-30 atoms thick and maybe 80-100 across. There ain't much juice left to squeeze in terms of size but perhaps we'll figure out how to reduce leakage (heat).

Physics hates the very large and the very small. The large get the tyranny of volume scaling + general relativity. The small have their entire concept of reality smashed by quantum mechanics.

Prior to LLMs there was some minimal effort required to snipe someone and possibly your reputation was attached otherwise there would have been no point in publishing to begin with.

Now anyone with a few dollars can do it, many who don't have a reputation to worry about.

It's the same problem as YouTube AI slop, AI-generated music, and everything else. Don't you dare tweet or blog about a video idea or hum a few bars from a song - within an hour 27 people will have posted AI slop rip-offs. There's something uniquely depressing about being beaten to the punch by a thief that doesn't deliver the same soul-crushing impact as having someone copy you after the fact.

They'd have needed the mental bandwidth to notice on the EICAS display that the spoilers weren't deploying to even know to manually trigger them. Given how far behind they were noticing that would be highly unlikely. Under high stress you revert to your training or habits and manual spoiler deployment is just not something you need to do.

Their bank angle (for lack of a better term) was probably not enough to notice unless they had the mental bandwidth to pay close attention.

This is a common thing in these sorts of accidents: if you have the mental bandwidth to deal with these problems you'd have been properly setting yourself up for landing 30 minutes before you started and you wouldn't be in this mess to begin with.

This also represents a safety margin during normal flights: If everything is done by-the-book and going well then you are calm, collected, and you have the mental resources available to do quick instrument scans after touching down and potentially notice spoilers did not deploy - and you have the extra runway to account for the few seconds of delay it takes to notice then manually deploy them.

To be stabilized you must be lined up laterally (eg localizer captured) and on glide slope (vertically aligned). Airspeed must be Vref +/-10, final landing configuration (flaps, gear, etc), sink rate correct (<1000fpm), power set (not idle), and all checklists complete.

Typically the pilot monitoring should make a callout for this (eg "Approach") and the pilot flying responds ("stabilized" or "unstable, going around").

The "minimums" callout is to verify that you can see the runway sufficiently to land. The exact requirements vary - a Category I ILS you can go down to 200ft without seeing the runway. A Category IIIb ILS is 50ft. Technically Category IIIc allows 0 so has no decision height and requires no "minimums" callout but I don't know of anyone crazy enough to allow this - you wouldn't be able to taxi off the runway with zero visibility.

The stabilized approach gate is a performance & configuration check. Is the aircraft configured for landing and on the correct performance profile with the expected automatic guidance?

Minimums callout (DA/MDA) is a visibility check. Can we see the runway to whatever degree our approach requires?

btw: If you are curious the various ILS categories impose performance and redundancy requirements on both aircraft and the airport ground equipment. Things like backup transmitters, self-monitoring to catch drift, battery backups, etc.

YMMV though, I'm not an ATP.

Part 121 Cargo and Part 135 Charter are ... another story. Quality of training and procedures vary a lot. Some are run about as well as the Part 121 Passenger operations. At others a go-around is a career-limiting move.

I do not know how 21Air operated - leave that to the NTSB to discover - but it would not surprise me whatsoever if Amazon leans heavily on their contracted ACMIs and yanks contracts for delays. A go-around means an extra 15 minutes (give or take) and Amazon cuts their logistics schedules razor thin so they don't have to pay for extra labor. A 15-20 minute delay in unloading the airplane can mean the sorting center staff and delivery trucks sit idle (though it wouldn't surprise me if they just stiff the delivery contractors for the wasted time).

This isn't necessarily done with ill intent - the middle managers at Amazon are just chasing their numbers and an operator that prioritizes safety will inevitably have worse on-time performance. As is often the case these days outsourcing / contracting can end up being a race to the bottom even if no one intends it that way. FedEx Ground is all contractors who bid on routes. If you ever wonder why the driver throws your package or leaves it in your driveway: they had to be the lowest bidder to get the contract and that means X deliveries per hour or they lose money. It just so happens that with air cargo operations this kind of pressure can easily lead to an accident.