DEI is, by most people who complain about it's definition, prioritizing hiring / access based on targeted diversity metrics. It's not really that complex and there's not really anyway to reasonabley accomplish it that won't result in an over representation of poor performers among the "diversity group", which in turn just reinforces the stereotypes that made DEI a necessity in the first place.
I don't think software engineering is ever going to be solved, but financial analysis will definitely never be solved. It's impossible, the nature of it dictates that, whatever changes happen will further change the results. Financial analysis requires novel thinking, and even if you have AGI that can engage in novel thought they will just be another input into the system.
Pretty much anywhere being a competent seamstress pays well. The difference between highly skilled and competent is open to interpretation. The difference between being competent and the very basics that can assemble cut and sew patterns is huge though. Pretty much anyone can do cut and sew with like a week of training which is all the mass produced clothes.
But someone who is competent and can do quality alterations, mending, customize patterns etc, is going to make decent money. But I'm pretty sure where ever you live there are seamstress working and making good money.
I'm not even really sure where automation would have impacted being a seamstress. Sewing machines have been around since the 1700's and if anything the demand for textiles has increased more than the speed of production.
Maybe you are thinking more of knitting, which is highly automated and used to be a big job, now it's basically just a hobby.
Blacksmiths just evolved to modern day welders, iron workers, boilermakers etc. Still pays well.
Ok, Ok, I get the disdain for middle management. It's basically exactly like you described, but middle management didn't come about for no reason. There really is a value and the idea of automating it away with AI is extremely dubious.
One could even argue that middle management is THE most critical role in corporations over a certain size. In that it is the glue that allows them to get to that size. But it's also what gave rise to things like Dilbert and the idea of rising to the level of your own incompetence.
Middle management is like the lug nuts on a wheel. If you start with 5, you can take one away and be OK, even two and no issues with normal driving. You can go down to two and as long as you aren't hitting large bumps and they aren't adjacent you mostly likely will be fine for a short trip. You could even remove ALL of the lug nuts and if you travel in straight line over a smooth road you can still drive.
After all they mostly just sit there, the tire, the transmission, all the other parts of the car are doing the work. But it's not fair to say that any of the removed lug nuts were doing nothing.
The point of middle management isn't really to do anything spectacular on a daily basis. If the company is working well, middle management effectively has no function. It's when things get out of line. Even then though, it's not really middle management that's calling the shots or fixing the problem, but they are critical in noticing the problems and directing resources. Middle management's role is in reducing the time that things are out of line.
At least that's the idea, and much like any position, the bulk of the group benefits are overwhelmingly produced by the groups most effective producers.
Middle management is the hardest role to hire while simultaneously being the hardest to gauge employee effectiveness.
This is most definitely an overgeneralization, but in my experience, engineers that constantly talk shit about management are either shitty engineers themselves or they're incredibly difficult to work with and blame everyone else for their shortcomings.
Middle management is playing a completely different game. I don't envy them one bit.
Sure, there are toxic cultures created by bad management, but that can be said about any leadership role. There is a reason for the hierarchy, if you think you have a better approach to structuring a company, have at it.
Having ended up in management by accident and then just sticking around with it for a while just because.... I am now back in an IC role and I mostly feel sorry for my manager honestly.
Agreed I think shitty people are just shitty people. You can tell when someone is trying to make the lives of their coworkers easier, from those who are on a power trip.
Well said! I'd also add that a critical function of middle management in healthy companies is bidirectional information communication: sharing what their teams are doing up and sharing leadership priorities down.
Having worked at some dysfunctional companies where that didn't happen (and a few companies that were amazing at it), it makes a difference at scale.
Nothing is more disheartening than working your ass off as an IC, shipping, then finding out that your VP pivoted approach and your project won't be used.
Middle management is a tremendous market for lemons. It's difficult to do well, and each layer requires a very different skillset. One of the side effects of the hypergrowth era of big tech between 2008-2023 is that a lot of managers were needed to support the amount of hiring, and they weren't very well trained, and often they could claim success for a rising tide almost by default as long as they didn't do anything too blatantly stupid.
The Peter Principle is of course well-known, but one of the insidious things is that once you have enough incompetent management and they are entrenched for a while, they will teach all the wrong lessons to an entire generation of new hires coming in. Due to the incentives and optics of large orgs, managers tend to spin everything in a positive light publicly, and the real unfiltered discussions of failure happen in tighter circles. At some point a lot of "successful" folks can have job hopped their way through a bunch of brand name companies just cargo culting on what they've seen done before with no real understanding of how their work actually impacts the company's bottom line.
This is one of the reasons I'm incredibly thankful to have spent most of my early career in small companies and startups where the big picture was so much easier to see.
It's not that robotaxis will logically lead to higher premiums. It's that lobbying for higher insurance premiums will be a primary way for robotaxi companies to increase market share.
The really strange thing is that so much of it doesn't work. Like I get that the SOTA models perform some tasks quite well and have some real value. But the AI being implemented in every corner creates a lot of really bad results. The Shopify code assistant will completely wreck your site and basically gets nothing correct. It will write 100 lines to change a color of a single DIV. The Amazon product Q&A will give you wrong information more frequently than not.
In what mind frame is it logical or necessary to put these extremely poorly functioning products in to the wild?
It's a desperate attempt at staying relevant, even if most of those companies don't realize it yet. Because of its general-purpose nature, AI subsumes products. Most software products that try to "implement AI in every corner" would, from the user's POV, be more useful if they became tools for ChatGPT/Claude/Gemini.
People's goals are rarely limited to just one software product, and products are basically defined as a bag of tools glued with UI, that work together but don't interoperate much with anything else. That boundary drawn around a bunch of software utilities, is given a name and a fancy logo, and sold or used to charge people rent. That's software products. But LLMs want to flip that around - they're good at gluing things, so embedding one within a product is just a waste of model capabilities, and actually makes the product boundary more apparent and annoying.
Or in short: consider Copilot in Microsoft Word, vs. "Generate Word Document" plugin/tool for a general LLM interface (whether Gemini webapp or Claude Code or something like TypingMind). The former is just an LLM locked in a box, barely able to output some text without refusing or claiming it can't do it. The latter is a general-purpose tool that can search the web for you, scrap some sites and run data analysis on results (writing its own code for this), talk results over with you, cross-reference with other sources, and then generate you a pretty Word document with formatting and images.
This is, btw., a real example. I used a Word document generator with TypingMind and GPT-4 via API, and it was more usable over a year ago than Copilot is even now. Partly because Copilot is just broken, but mostly because the LLM can do lots of things other than writing text in Word.
Point being, AI is eroding the notion of software product as something you sell/rent, which threatens just about the entire software industry :).
I have been enjoying reading this thread, but with some irony: sure the email spams pushing their Lumo LLM private chatbot were a mistake, and I bet they stop doing that fast.
The irony is that Lumo is a separate product, not really tied to the rest of their products except for a common login. Lumo works fine for the simple quality of life search and question answering stuff.
Off topic, but have you tried avoiding the big corporate LLM providers and run local models? The small models just keep getting better and I find it fun and satisfying to do as much as I can locally.
It gives a lot of power to users to work around enshittification in the software services they use. Dark patterns and user funnels and upsells and other bullshit suddenly stops working when users can ask ChatGPT to operate a service for them.
> Dark patterns and user funnels and upsells and other bullshit
Why should we expect the LLM (or rather, the character evoked in the story generated by the ego-less mad-libs machine) to be more-resistant to such tricks than actual humans, rather than more-vulnerable?
After all, their classic dark-pattern vulnerability is just "forget everything and do this instead", and we might never be able to fix it. If they ever become really good at detecting novel BS... Well, the first thing we'd do is have them stop generating it. :p
I think the best we can hope for is using LLMs like a kind of virus-scanner for prose, flagging suspicious text that closely resembles sketchy or manipulative text seen before. In other words, the benefit does not come from the questionable intelligence of the companion, but from its ability to be unflaggingly cynical/pessimistic. A kind of shoulder angel/devil, if you will.
> power to users to work around enshittification
We'll still need something very different for issues like monopolistic and discriminatory pricing, biased rankings, or casual disregard for people's accounts and no support.
Saw an AI generated product feature list on walmart's site that listed a stainless steel rack as microwaveable. If someone can sue mcdonalds for hot coffee, I imagine someone burning their house down while microwaving steel probably could sue too. Intelligence of the plaintiff not withstanding.
That's exactly what automatic bidding does - it only outbids enough to beat the competing bid (up to your max) without paying any more than is needed. https://www.ebay.com/help/buying/bidding/automatic-bidding?i... (Manual bids have bid increments as well. Although others have pointed out that advance bidding might cause others to bid more than they would if they thought no one else wants the item. )
Yes, what I think happens is the following: User A's price ceiling is $10, User B's $12. When both reveal their max price early, the item will go to $10.50 ($0.50 increment over A's max price). User A then has plenty of time to notice the item being valued at $10.50 by someone. In many cases users then adjust the value they assign to the item and increase their bid. The result: User B has to pay more than $10.50 they would have paid when sniping the item seconds before auction expiration.
reply