They are rerouting RAMs for consumers to enterprise for server build up - for higher margins I’m sure. MAG7 will happily pay more but poor consumers like us can’t - this is more bad news for us.
Wondering if we're going to have a situation in the future where we end up having to buy the hand-me-downs from industry after they're done with them (and thus kind of outdated tech)? Kind of seems like the days of building your own PC are numbered.
I prefer to just not think about it and blame charging my car. Also in Cali.
I don't know if I'd buy any of the servers or older computers but the internal components are pretty dang good. Stuff like hard drives, network cards, hbas, that's the real money saver right there.
Even the hand-me-down situation has been changing. I remember in the 2010s being able to shop at Weird Stuff in Sunnyvale, which was a treasure trove of old enterprise equipment, a heavenly place for a retrocomputing enthusiast like me. I even bought accessories for my NeXTstation there! The shop kept the good retro stuff in the back away from the general public for people who asked kindly about retro gear :). I saw some 1980s and 1990s classic Macs, too, in the restricted area.
Unfortunately Weird Stuff is a thing of the past, though this had less to do with a reduced supply of surplus and more to do with Silicon Valley’s high real estate prices. Thankfully there are still good stuff to be found at e-waste recyclers, but if more companies are relying on the cloud, and if more devices are integrated boards with everything soldered on (such as most modern Apple hardware), the hand-me-downs of the future are going to be harder to work on than today’s hand-me-downs. I’m just an average software guy whose hardware experience is limited to a graduate-level computer architecture course and building PCs. I can talk about caching and branch prediction, but I’ve never picked up a soldering iron in my life; I’m no Louis Rossmann.
Hyperscalers and even "smaller" shops by comparison, such as CoreWeave, TerraWulf, Lambda.ai.
So yes, enterprise customers doing AI data centre buildouts. They are going all out for them at the expense of their consumer business due to supply constraints.
I don't see this situation changing for many years to come. Would indirectly affect the cost of any electronics that has storage or memory on it. Would be interesting to see how Samsung plays this one out with their limited inventory - they make RAM + SSDs, use it themselves on their phones, laptops, etc. Also supply to consumer and business customers. Interesting times.
Your argument doesn’t make any sense. What does this have to do with supporting Arm chips? It’s not like AMD and Intel are waging a war against Valve. If anything Steam helps them by strengthening the PC gaming market, leading to higher CPU/GPU sales.
Slowly getting their stuff independent of wintel gives a lot of flexibility. And the big gaming market's on phones / tablets. A steam controller could find itself paired to an iPad running steam in a year or two.
The only play I see here is a legitimate Valve console to take on XBOX and Sony. Plus Arm on a Steam Deck would improve the battery life considerably (assuming they are able to integrate with some powerful GPU solution).
I’ve always liked the idea of using Wolfram / Mathematica for exploratory work (mainly statistics and data science) and found it to be too academic for my taste. Not as simple as using say, pandas, where I can rely on editor autocomplete to help me figure out what I need. It’s a result of their functional design choice but it forces the user to know what they need. I have poor working memory and “let’s figure it out as we do it” works best for me. Wolfram lang is not good fit for that IMO.
AI models are getting close to delivering on the advantage it holds - like solid visualizations and good mathematics to programming translatability. In fact, I think their “engine” with a multi-modal AI input + MCP, would be the best of both worlds and may help push their adoption. Or perhaps even a copilot type experience in their IDE. When I look at their site now, it looks practically unchanged from 5 years ago - so I’m a little taken aback given Dr. Wolfram’s initial enthusiasm around LLMs, seeing a lack of any significant AI feature adoption.
Tbh missing a quote around a path is the most human mistake I can think of. The real issue here is you never know with a 100% certainty what Gemini 3 personality you’re gonna get. Is it going to be the pedantic expert or Mr. Bean (aka Butterfingers).
Though they will never admit it and use weasel language to deny like “we never use a different model when demand is high”, it was painfully obvious that ChatGPT etc was dumbed down during peak hours early on. I assume their legal team decided routing queries to a more quantized version of the same model technically didn’t constitute a different model.
There was also the noticeable laziness factor where given the same prompt throughout the day, only during certain peak usage hours would it tell you how to do something versus doing it itself.
I’ve noticed Gemini at some points will just repeat a question back to you as if it’s answer, or refused to look at external info.
Gemini is weird and I’m not suggesting it’s due to ingenuity on Google’s behalf. This might be the result of genuine limitations of the current architecture (or by design? Read on).
Here’s what I’ve noticed with Gemini 3. Often it repeats itself with 80% of the same text with the last couple of lines being different. And I mean it repeat these paragraphs 5-6 times. Truly bizarre.
From all that almost GPT-2 quality text, it’s able to derive genuinely useful insights and coherent explanations in the final text. Some kind of multi-head parallel processing + voting mechanism? Evolution of MoE? I don’t know. But in a way this fits the mental model of massive processing at Google where a single super cluster can drive 9,000+ connected TPUs. Anyone who knows more, care to share? Genuinely interested.
I get this too. I’ve had it apologize for repeating something verbatim, then proceed to do it again word for word despite my asking for clarification or pointing out that it’s incorrect and not actually searching the web like I requested. Over and over and over until some bit flips and it finally actually gives the information requested.
The example that stands out most clearly is that I asked it how to turn the fog lights on in my rental vehicle by giving it the exact year, make, model, etc. For 6-8 replies in a row it gave the exact same answer about it being a (non-existent) button on the dash. Then finally something clicked, it searched the Internet, and accurately said that it was a twistable collar midway down the turn signal stalk.
The big ones are also diesel, run at different RPM. ~3500 RPM For the smalle gas ones and around 1800 RPM for the big diesel ones. The noise characteristcs are also different. The big diesels produce more low frequency noise, whereas the small gas generators are literally lawn mower / leaf blower engines and sound the same.
The enclosures of the big generators remove mainly lower frequency sound, making them perceptually even quieter.
It’s crazy that the empathetic AI I sometimes share my life’s mundane problems with wrote this. It understands such high level maths and speaks in a lexicon with words I can’t even pronounce. Incredible versatility. If this isn’t AGI (or even super intelligence) then what is?
Sure it suffers from amnesia and cannot get basic things right sometimes, but one is a design limitation that can be overcome and the other a possible problem caused by training (we’re discovering that overt focus during training on adherence to prompt somewhat lobotomizes their general competence).
As for the first: Everything points towards the problem NOT being able to be overcome with current architectures.
For the 2nd: We are completely 100% sure this cannot be solved. This is not a training issue. This is the case for any statistical machine. No amount of training can ever solve this.
This is a poor take on what LLM's actually are. They don't understand in any real sense.
If you had enough paper and ink and the patience to go through it all you could take all the training data and manually step through and train the same model. Then once you have trained the model you could use more pen and paper to step through the correct prompts to arrive at the answer. All of this would be a completely mechanical process.
> but not some fresh graduates who can work 80 hours per week and only demand half of the salary
Cause garbage in, gets garbage out. With AI models being all the more rage in the coming years, unexperienced hires will prove many times more costly. (10x garbage with agents).
So companies are going to concentrate their worker base even more with experienced folks. They need fewer of them. Yes. But quality matters more than ever.
I really feel bad for the new graduates. For no fault of their own, the bar went up so high. Unless they’re a child prodigy doing some coding projects on the side since the age of 10 - no one will hire them. So how will they ever gain the experience they need?
Maybe, just maybe, we’ll see a reinvention of coding schools - that will now focus on fundamental and industry knowledge - imparted by other veterans, instead of teaching applied skills.
I agree with you, but I forgot to mention that in the original reply I meant to say that "After the economy turns around, there is no point to hire me, an older guy with maybe a couple of gap years, who worked as a Uber driver for the last two years and can't leetcode".
But yeah, new graduates is going to suffer anyway.
And I'm scared of the collapse of the existing world order. Maybe we won't see a turn around for many years if it does collapse -- and we are already seeing many cracks on it.
New folks will never be hired. RIP to the CS degree.
Old staff will be exited. Especially senior and mid level management.
If you lose your job, you won't get the same comp again. The days of $500K TC are long behind us.
It's the era of downsizing and outsourcing while blaming AI.
None of this has anything to do with AI. That's just a scapegoat.
Google and Amazon are culling entire US teams and rebuilding them in Asia where the cost of labor is significantly lower.
The best thing ICs can do is fight for big tech monopolies to be broken up. (Call your reps leading up to the midterms.) If several members of the Mag 7 are broken up into smaller companies, that'll inject tons of energy back into the ecosystem and enable the wheels of competition and employment.
Bonus - if big conglomerates are fighting to pick up the pieces of a Ma Bell style dismantlement, they won't have time to manage teams 12 hours away.
Nothing against our colleagues in Asia. They're brilliant. But American companies built with American labor shouldn't shut us out in the cold while they reach record profits and continue to hollow out entirely new industries simply by outstretching their arms.
I’ve been told for 20 years that in 5 years my job is going to be offshored. If they could have they would have long ago.
My theory: We had a crazy bubble of hiring during zero rate interest. We are living through a nasty correction. AI is moving the needle too, but it’s mostly being used as a scapegoat to save face and explain away cleaning up failed ZRIP yolo plans that didn’t pan out.
We’ve also haven’t had a serious recession since 2009. It feels like it’s only a matter of time :(
> I’ve been told for 20 years that in 5 years my job is going to be offshored. If they could have they would have long ago.
"This time it's different."
20 years ago China and India had a nascent tech industry. Now they're booming.
Talented folks all over the world - Asia, Latin America, and elsewhere - are working on hard problems.
> We had a crazy bubble of hiring during zero rate interest.
We did. This has had a tremendous impact, no doubt. But by the same coin, ZIRP has had half a decade to unwind at this point. There's other stuff going on. Tariffs, continued inflation, etc.
We're not the only industry offshoring. Hollywood has moved a lionshare of production overseas in the last 4 years. Graphics design and marketing... It's being shipped out at volume.
I was told that once video conferencing got good and internet and infrastructure became better in other places, "this time it will be different."
I was told once universities in other countries started pumping out a pool of great candidates, expats who worked for FAANGs in the US would go home to found their own companies using that pool, and those companies would take over the world. "this time it's different."
I was told during covid once everyone was remote, why would people not just hire the cheapest remote workers going forward? "This time it's different."
Don't get me wrong, I absolutely have seen more and more offshoring over time, but there's a huge inertia behind the US tech industry that's hard to change. The VC / startup ecosystem and all of it's resources have huge Bay Area inertia - it mostly hasn't even spread to the rest of the US, let alone the rest of the world, despite the cost of living and constrained talent pool in the Bay Area. There's something about getting a bunch of people with the same mindset in one spot and having them know each other, socialize with each other, make friends and networks with each other that still matters. Founders tend to build off the people and connections they know and are connected with personally.
I'm hoping it will take long enough to change for me to finish my career. We'll see. This time really may be different :).
EDIT: p.s. Agree totally it's way to complex to tell what's actually happening. The _impact_ of the end of ZRIP, the rise of AI, major tax changes on R&D amortization, and US tariffs pretty much landed at the same time, so who knows?
The difference is, offshoring in the 2000s was largely private sector driven with minimal subsidizes and investment promotion programs lead by countries and their local governments.
On the other hand, in the 2020s, India, Israel, most Eastern European states, Ireland, Costa Rica, and a couple others have launched industrial promotion subsidizes for software offshoring - often providing US$10k-30k per head in federal and local subsidizes along with subsidized office space and real estate and tax windows.
That along with the internal frictions of async work largely being ironed out due to the COVID remote work period along with an exodus of mid-level managers on work visas during the early pandemic layoffs which had an outsized impact on Indian, Chinese, and Eastern European techies in the US made offshoring much more cost competitive and effective than it was 25 years ago.
Putting your head in the sand saying it's no big deal is honestly very stupid if you are hoping to maintain your career for the next 5-10 years in any white collar job.
And it's only going to get even more competitive now that the Indian government is enacting labor reform laws to align Indian labor laws with China's [0], making it even more cost effective for businesses to offshore by reducing regulatory overhead [1].
> On the other hand, in the 2020s, India, Israel, most Eastern European states, Ireland, Costa Rica, and a couple others have launched industrial promotion subsidizes for software offshoring - often providing US$10k-30k per head in federal and local subsidizes along with subsidized office space and real estate and tax windows.
This isn't true either in India or most of eastern europe.
Maybe you are confusing PLI for manufacturing? Altough even that's not on per head basis.
> This isn't true either in India or most of eastern europe.
It is.
Wage arbitrage doesn't move the needle for offshoring once operating costs come to play, and outsourcing companies like EPAM, WITCH, and others juiced their margins by padding heavily, which further reduced the cost competitiveness of outsourcing without subsidies.
Czechia [0], individual Volvodships along with the federal government in Poland [1], state+center in India [2][3], Ireland [4], Romania [5], and others [6] dated list from KPMG which doesn't include state and local incentives) are all providing subsidies for GCCs now which include a payroll/per-head incentive depending on the amount spent in FDI, along with added additional subsidies per industry (eg. Life sciences GCCs get additional sets of subsidies versus a generic software GCC versus a VFX GCC).
The US has some of the weakest R&D tax incentives globally [6], with no payroll or financing incentives - only Vietnam, Philippines, Peru, and PNG are stingier, which has been a major role for why GCC expansion has been rapidly growing for the past few years.
That said, these incentives are primarily targeted at large employers becuase if you cannot provide at the minimum dozens of jobs, then the cost cannot be recouped over the long term by most subsidies. So mom-and-pop 3 person consultancies are ignored because in most cases they are parasites and large firms interested in opening large dedicated headcount offices are incentivized.
I got my first programming job in ~98 while still in college. I had family members telling me then that programming was a dead end and was all going to be outsourced. I lived through .com, GFC, etc... This does feel more like a reversion to the mean at this moment rather than some crash...yet. I feel for the people who only have known a job market that was easy to step into and paid great salaries because they don't know anything else. It's a lot like the people who think they are great stock pickers because they've only been investing in the greatest bull market we've ever seen.
When I came into the job market the rule of thumb was it would take 1 month/10k of salary to find a new job. Over time that moved to 1 month/20k of salary or so. Even then, someone making a FAANG type salary should be prepared to look for a ~year for a new job matching that salary. Being able to bounce from job to job while getting big raises along the way was the exception, and ZIRP only exacerbated it.
You're zooming out and considering this negative sentiment with similar times in the past. I think that's wise. I think we should keep zooming out to other industries. Imagine you're an engineer for GM in Detroit in the 70s - would you consider the mean to be your contemporary middle-class lifestyle, or what it is in 2025? Similar for steel and semiconductors.
It goes for other places, too. Is the US's financial strength of today its mean, or is it where the UK was pre-Suez Crisis? Where Japan was in the 80s?
Let’s hypothetically say we’re all doomed. Say our jobs are going the way of manufacturing in the 70s-80s. What’s the play then?
If I was a new college grad I’d stay away from programming, but that’s been true for a while regardless of offshoring, the job market is just too soft until managers figure out they are killing their senior engineer pipeline and go back to investing in people.
What about the people already in industry? What’s our play?
Live under your means and save as much as possible? Already doing it.
Learn a new trade? Does not feel realistic while working a demanding full time job already, but if things get bad enough, sure.
Use the political apparatus to protect my employment? The system is built to prevent me from doing that. Fighting the system very well could put my employment at risk, which defeats the whole “get what I can while I still can” plan, if I assume doom and gloom on the horizon. I’m also unlikely to actually change anything by taking that risk, so the ROI is horrible.
Is there some other outcome or plan of action here I’m not seeing?
>Live under your means and save as much as possible?
which, while obvious, isn't being done by all of us.
A part of me gets angry that collective action was so unpopular thanks to the view that it dragged down those who could excel individually. Every time I see software people act powerless in front of these steamrolling, enormous tech giants that control every facet of our lives, I think about how much power we had - and are on the verge of giving up.
I also try to confront the future, rather than turn a blind eye to it. Can I be happy and find self-actualisation without this identity and financial status? That's a question everyone should think about regardless of what happens.
I love the idea of FIRE as a life goal and driving financial strategy. The core principal is you save enough money up that the dividends from your investment returns (the FI part of FIRE) is enough to live off forever.
If you hit FIRE, awesome, you’re free from ever caring about offshoring or RTO or AI or whatever again.
If you don’t hit it, you’re sitting on a pile of money when a rainy day comes.
Adding my voice to sibling comments, this is from European experience, I have had several times been dumped from consulting projects, and having to do competence transfer to the offshoring team that would take over our team roles.
This is not the picture I'm seeing on the ground. AI is eliminating certain classes of junior software positions. (Roughly: jobs where explaining a task to junior engineer is more work than asking Cursor/Claude Code/Codex to do it.) Junior engineers can fight back against this by
a) getting really good at clarifying requirements
b) learning quickly, so their work quality is eventually higher than Cursor can work out in one shot.
This is also a pressure against hiring teams overseas: when the bottleneck is communication + taste, not raw implementation cycles, you'd rather have a small local team. And it's a pressure for high TC, because individuals now have much more leverage, although they need to master more skills to take advantage.
Many juniors can't even meet with a human interviewer. There's no point maximizing for interviews that never come. That's the issue.
>This is also a pressure against hiring teams overseas:
This seems to agree with the issue. a team of 100 becomes a team of 5 locals and 95 outsourced work. Maybe those 5 managers are better off, but we're still reducing the local workforce by 95%.
And I doubt the conditions of the remaining 5 are better than pre-outsourcing. You can't out-compensate burnout and QoL. Gen Z in particular seems to really be pushing against this mentality, so this strategy is limited in time even if it's working on Gen X/Millenials.
Being real good does not change the fact, that one is cost factor and at the end only a row in payroll spreadsheet. Junior with low salary and low compensation during layoff -> priority departure.
Having in couple hours unannounced meeting. My boss told me over private channel, that he just got fired. It’s very interesting and the home mortgage does not really help today. I was really good. Better than expected and accomplished few optional projects. Looks like it didn’t help again.
> learning quickly, so their work quality is eventually higher than Cursor can work out in one shot.
This sounds almost word for word like The Onion’s classic: Secretary Of Labor Assures Nation There Still Plenty Of Jobs For Americans Willing To Outwork Robots
Wow. That is ... too painful and true to life to really be funny at this point. But, ok, still funny.
To be fair, I meant something a little different -- something like -- learn how to be a robot priest who can get it to follow the desperate prayers of humans. And, like, how to unstick the robot arm when it accidentally punches through a wall. Etc.
Not that that is particularly comforting, in an existential sense. Maybe buys you a couple years till you have to pivot again.
That was the bottleneck in the industry when it was in growth phase, it's a mature sector now and it's all about efficiency and profit now. Speed to market and product iteration speed isn't the most important thing anymore, there's not a lot of innovation taking place. Outside the actual novel AI specific companies out there, of course, there are a few other spots of growth and exceptional companies but largely the kings have been crowned.
Show of hands for anyone seeing AI replacing juniors (and I assume also backfilling employees).
I'm genuinely curious.
I've heard this argued the other way too. Seen it firsthand.
Fwiw, we've had good engineers switch to vibe coding and it's ruined their output.
From really solid systems to unmaintainable flocks of seagulls - nested if statements ten levels deep with no thought or care. From good engineers that are just dialing it in now.
We've had good engineers use vibe coding to save to time to work on their side hustles. Then go on to try to raise money for AI products.
I lead cloud consulting projects as a staff consultant specializing in application development.
I use to need myself to lead the project, customer management, design work and some development. I would add usually another developer to do some of the grunt work coding and usually a cloud architect to take care of infrastructure as code, security, etc.
Not that I wasn’t knowledgeable enough to do it all myself, I just didn’t have time. GenAI can definitely do CloudFormation, Terraform or the AWS CDK (ie using a high level language like Typescript instead of Yaml) and can do the code where I really don’t need two other resources or deal with the detailed requirements and coordination.
Before the pearl clutching starts about my not knowing how to code without AI. I’ve been coding consistently since 1986 when I was a hobbyist assembly language coder.
> We've had good engineers use vibe coding to save to time to work on their side hustles. Then go on to try to raise money for AI products.
this is honestly what i think is going to happen as everyone is still figuring this out and why for large companies i think they are walking into a giant ass trap and the execs can't or don't care to see it as long as it boosts the stock.
most people (read: MOST) working at X corp aren't going to be using AI to accelerate their work, they are going to "phone it in" like you said or buy themselves more time to do other stuff. i think with the onslaught of vibe and slop most people don't have a vested interest to go above and beyond to ensure quality as long as it works, even experienced devs. that's honestly probably what i'd be tempted to do.
where we see the most exponential gains will be people who are paid enough to care, smaller teams that have a direct vested interest in long term success (founding teams, etc.) or people paid to just clean up a mess.
hiring and the landscape will be super interesting to see in the next year or so. probably team structure will change drastically as well.
Yeah pretty much. Engineers are going to be at a crossroad where they either turn to the government to finally build in some proper labor laws and other obvious controls (how about re-banning stock buybacks?) or go out to the Wild West and hope they idea can sustain their livelihood.
Given the vibes of the community here: I guess I'll look for a Mad Max mask (I'll ofc keep performing my civic duties, though).
A big tech breakup needn't be anti-capitalist. In fact, it might be the most pro-capitalist move.
If you're an entrepreneur or VC, you want big tech broken up because they can put serious price pressure on your exit.
Trillion dollar companies can easily spin up a team to copy you, with no incentive to stay alive. They can threaten you with all kinds of leverage - access to customers, patents, legislators. They can give you an ultimatum to sell for cheap, go to your competitor, etc.
Their scale and reach is additional unexpected gravity on your delta V.
Capitalism is supposed to be hard. It isn't supposed to support invasive species that can graze anywhere they please and kill ecosystems of diversity and innovation. These mega conglomerates can just throw themselves into markets using unrelated business unit profit and suffocate real companies.
Breaking up Google and Amazon would be good for everyone, perhaps even shareholders and ICs at those companies themselves if value is unlocked. Let alone all of the other companies and entrepreneurs in the market.
I think it depends on which kind of entrepreneur you're aiming to be. VC breakups are amazing if your goal is to box in and become a market competitor. But as of the last decade or so there's been plenty of "incubators" to take into account. startups whose goal is instead to be sold off to some major company and get their payday that way. Those kinds of models would deteriorate, and are likely what want to prop Big Tech up.
I do hope we have more genuine competitors fighting out there for breakupps. But it's hard to say these days.
> But as of the last decade or so there's been plenty of "incubators" to take into account. startups whose goal is instead to be sold off to some major company and get their payday that way. Those kinds of models would deteriorate,
Good, because this is an extremely toxic and damaging version of capitalism
Every company should be seeking to stand on its own, not become assimilated by the borg
This is pretend boogie man. Banning buybacks will not automatically make that money flow into hiring or salaries. Companies are not charities, they exist to make a return. If hiring people and/or paying more will generate a larger return than giving the money back to shareholders either through buybacks or dividends, then companies will do that.
AI is now giving companies something to do with excess cash that could generate better returns (shareholders believe so) and buybacks are being pushed out as money goes elsewhere[1].
>Banning buybacks will not automatically make that money flow into hiring or salaries.
Nope, but that's what trends show us from the decades between its ban and bans being lifted. All I know is that companies flowing money back into itself and having executives shift in and out every few years clearly hasn't worked.
It's just one stepping stone to make sure companies have skin in the game again.
You can argue dividends but that means the money gets taxed quicker, so that also helps the people.
>AI is now giving companies something to do with excess cash that could generate better returns
Sure, for now. I think that problem will fix itself sooner than later, so I'm not too concerned about that. Trends come and go.
Stock buybacks are designed to let the shareholder see the same upside, but decide when to take the taxable event. Long term gains are also preferable to non-qualified dividends.
Even worse for non-American stockholders of American companies - the IRS charges a 30% foreign withholding tax on dividends. If you ban stock buybacks in favour of dividends, it’s a big tax increase on foreigners, so US stocks lose a whole pile of value for American stockholders when foreigners dump American equities until the ROI equalizes. (Roughly 20% of US equities are foreign-owned.)
Dividends don't grow the stock as quickly. They can and will do that, but the goal is to change the incentive structure back to long term growth and not "stock buyback and dip from company in a year".
There is no evidence to indicate buybacks reduce long term incentives for executives compared to dividends.
If anything, all the businesses with the most long term growth have done the most buybacks because they are paying the employees in stock, which employees gladly accept because they bet the business will have long term growth.
And executive compensation is not vested until business targets are met a few years in the future.
> There is no evidence to indicate buybacks reduce long term incentives for executives compared to dividends.
If I am bonused on earnings per share, and I have a button to increase earnings per share mechanically (without needing to increase revenue or decrease costs), why wouldn't I push that button?
Can you share some evidence around your statement? i.e. "There is no evidence to indicate buybacks reduce long term incentives for executives compared to dividends."
> If I am bonused on earnings per share, and I have a button to increase earnings per share mechanically (without needing to increase revenue or decrease costs), why wouldn't I push that button?
Examining fantastical scenarios is a waste of time. No one has that button, there is a whole board of directors that votes on these things, and again, compensation is staggered over various performance targets staggered over a number of years. The proxy reports detailing these are easily accessible with an online search.
> Can you share some evidence around your statement? i.e. "There is no evidence to indicate buybacks reduce long term incentives for executives compared to dividends."
Reality. The businesses that have the best long term performance over the previous decades are the ones that have done the most buybacks, hence buybacks do not cause short term-ism. It’s just as easy as a business cutting expenses now to juice dividends in the near term, we’ve seen it time and time again with businesses that sacrifice quality and innovation in the short term which eventually cede ground to new businesses.
> Reality. The businesses that have the best long term performance over the previous decades are the ones that have done the most buybacks, hence buybacks do not cause short term-ism. It’s just as easy as a business cutting expenses now to juice dividends in the near term, we’ve seen it time and time again with businesses that sacrifice quality and innovation in the short term which eventually cede ground to new businesses.
So no source(s) then?
Your statement is just as un-evidenced as mine, so I'm uncertain why you seem to be so sure of your statement.
> Examining fantastical scenarios is a waste of time. No one has that button, there is a whole board of directors that votes on these things, and again, compensation is staggered over various performance targets staggered over a number of years. The proxy reports detailing these are easily accessible with an online search.
If I can reduce the number of shares, then I can increase EPS. Buybacks reduce the number of shares. The button clearly exists, and based on watching this happen it's clear that the button gets pressed pretty often.
> The businesses that have the best long term performance over the previous decades are the ones that have done the most buybacks, hence buybacks do not cause short term-ism
This is mixing up cause and effect. If I have lots of free cash flow, then I can do buybacks which increases EPS. Lots of free cash flow is associated with better performance, which is perhaps why people believe this.
> New folks will never be hired. RIP to the CS degree.
We've just hired a couple of graduates, with the expectation that they are going to take some time to grow.
What I'm seeing right now is a huge influx of candidates from large companies that have zero skill. I'm not exaggerating, they can't code anything. And it's not just AI, they started working before ChatGPT came out.
Others in the industry are seeing the same and it's quite likely that your resume is getting lost.
One practical advice for resume writers from me. PLEASE, just don't put stuff like "Improved the API responsiveness by 23.123897%". Unless it's a crazy number like 100x.
We posted a job a year ago for a dev. We received terrible candidates, but still tried to fill it from the pool. 2/3 ghosted the interview and the other I'm not sure had ever done anything in iOS. I just pulled the job instead of wasting more time. I'm planning to post another job in the new year and I'm not looking forward to wading through the garbage.
> What I'm seeing right now is a huge influx of candidates from large companies that have zero skill. I'm not exaggerating, they can't code anything. And it's not just AI, they started working before ChatGPT came out.
This has been true in software for decades. From the very first time I was senior enough in my career to start conducting interviews on the "employer" side of the table, we've seen a huge number of candidates who literally (in the literal meaning of literally) could not code. Like you would ask them to write a for loop, and they froze up and couldn't do it, or just started talking, hoping we would move onto more "behavioral" questions. This has been pretty much a constant in the software industry for as long as I've been in it.
> Nothing against our colleagues in Asia. They're brilliant. But American companies built with American labor shouldn't shut us out in the cold while they reach record profits and continue to hollow out entirely new industries simply by outstretching their arms.
What makes you think people in Asia wouldn’t benefit from more competition in the market as well?
That said - I feel that advertisement based markets will always consolidate. There is too much of a benefit to having a single network which has the largest reach in terms of audience to show ads. This will always create incentives to consolidate over time.
Then again, why make the perfect the enemy of the good. Getting to more competition is a good step.
We gotta gather ourselves and remind companies why they once paid handsomely to not let potential disruptiors run rampant on the market. Long term new teams will form once productivity is valued again and not this giant incestuous GDP-maxmizing scheme.
I doubt things will recover to 2018 levels. Too many new software devs coming out each year, too much AI, too little big company growth once everyone already has an internet computer in their hands. The Wild West is over and now the digital economy has entered the boring phase.
The comparison to greenhouse gases doesn’t make sense. Corps pay a lot for developers right now because they get more value out of them than they cost. As long as that remains true, devs will be fine.
> I think we ought to be keeping people trained and employed
I never understood this sentiment. We don't have a massive manual weaving industry anymore, 95%+ of people used to be farmers in 1900. Tech comes and replaces humans, and the transition can be extremely painful especially for the people replaced, but ultimately it's better than keeping people artificially employed in obsolete jobs.
(I don't think SWE will be obsolete, but even in this case I'd rather switch careers)
Most deindustrialized regions in the West haven't recovered to full prosperity and are quite depressing to live in, sometimes even 30-40 years later: US Rust Belt, Wallonia in Belgium, the French North East, etc.
At a large enough scale, most people don't really move on, their lives are wrecked and they just suffer through them.
This is predicated on the myth that people can re-skill and move into new industries. Sure the former can happen and people can learn new things. But we're talking about an economy where there are no new industries. And an economy where you have to work in order to live.
What's a software developer in their 30s, 40s, and 50s supposed to re-skill into? Take on debt for the rest of their lives and re-skill into a profession (if they can even afford to take several years out of their lives to go back to school)? Into blue collar work along with the salary cut for which they might not have the physical capabilities for?
There's no social system for providing the necessities for living.
The other side of it is skill. Human societies have lost knowledge before. We've had to rediscover various aspects of metallurgy before. We could lose the ability to understand the technology we've made if we trust everything to the LLMs. There are already vibecoders who don't even be able to review the code that it generates for them because they're starting to lose the critical faculties and skills to understand it.
Or, more dystopian take... it won't matter. If software reliability continues to degrade in a normalized fashion, it won't matter. First mover advantages and networking effects will make it impossible for an outfit trafficking increased quality to ever get enough breaths to even compete.
Depends on your definition of “compete”. Compete for VC funding and continually increasing growth? You’re right. Compete as in stay profitable and have a future? Less clear-cut.
I believe you’re in the minority here. Perhaps your experience is different because of your skill set or the market you’re in. Anyone that I know personally who got laid off (in tech) took at least 6 months to find a job. I don’t know about anyone else but that to me is pretty brutal. More so as the people getting laid off are mid career, some with kids.
Edit: Add to the above that companies like Walmart are seeing an uptick in high wage earners becoming their customers, and McDonalds seeing a shrinking population of low-wage customers.
It’s easy to infer the rest from there. People who used to do well are cutting expenses and those who were already struggling are..I seriously don’t know what they’re doing. Where do you eat when you downgrade from McDonalds..Wendy’s? It’s a sad state of affairs.
You ask, "Where do you eat when you downgrade from McDonalds..Wendy’s? It’s a sad state of affairs." On the off chance that this isn't a joke, you need to know that eating out is very expensive in the US, even at McDonald's. According to the obviously highly credible https://mcdonalds-menu-prices.us/ a Quarter Pounder With Cheese costs US$7.99 now. I think home-cooked rice and lentils costs about US$0.20. Other similarly low-cost foods include polenta, homemade bread, homemade mayonnaise, zucchini, spaghetti, sunflower-seed cheese, homemade peanut butter, onions, potatoes, etc. Those numbers aren't even the same order of magnitude.
the advantage of fast-food over groceries, is that you don't have to worry about spoilage and waste. So the delta is probably less than you think. Now granted McD is an s-show, they are no longer the restaurant of the poor, You likely can get a better burger meal deal at a Chilis than a McD, as sad as that is.
Even if you waste half your groceries its still cheaper than eating out. And wasting that much is difficult to do, most staples will last weeks to years without risk of spoilage.
There are some fresh fruits and vegetables that are exceptions because they dont take well to refrigeration or freezing but really not much.
Flour comes in sacks, meat comes in cuts - we've a quarter lamb in the freezer, part of that in the fridge, and yeast and flour enough for bread for the next six months.
We shop cheap, like the family has done for the past 100+ years, much of our food comes from the garden - our excess gets swapped with others excess (we have a lot of fruit, we never buy eggs, they come from people that can be bothered to run chickens).
It's a bit of work, we save money by not going to a gym and our life expectancy and cancer survival rates are much better than, say, middle north America.
To cut mbfg a bit of slack here, your approach doesn't work in all situations. I admire your functioning community and supportive family and the fact that you've got time and space for things like gardening.
If people can't live like you do, it's probably because they've been placed on some kind of economic hamster wheel, and rather than figure out how to get a quarter of a lamb their better bet is to emigrate or to disrupt the system that's making McDonalds feel like a relevant factor in a survival equation by building the kind of community that you're describing. That's a big ask if you've never been part of such a thing (I know I haven't).
Sure, we live in an isolated area and have evolved through years of not even having a shop (well, I got to see one finally ~ 1980 or so). My father as young pre-teen helped support three younger siblings and a mother while his father was away at war by trapping rabbits and all that.
I had eighteen months as an isolated single parent with near zero support (long story) and had to stretch a marginal budget during that period. I've also travelled through the more remote parts of more than half of the 190+ countries across the planet, sorting logistics for food, fuel, et al along the way.
What I can pass on as hard earned lessons are that fast foods are rarely the cheapest or heathiest in the long term - if you can track down a large volume slow cooker in any garage sale or op shop on special you can keep a never ending stew on the roll by throwing in damn near anything, potatoes, celery, beans, carrots, bits of meat, swedes, etc.
It's hard to disrupt a system, difficult to break patterns and build communities and establish areas to grow food - but home cooking and stretching out food is something that can be found across the planet in the most unlikely places. Worth the effort to look for examples and make a few moves.
We're lucky to buy and prep all our food in bulk - it's more expensive on the infrequent shopping days, it's substantially cheaper over the course of year.
It's not something we need to do in current circumstances, it's a habit kept up in case it's ever needed and being frugal where possible means more to spend elsewhere.
This is just wrong. Beans and rice are more than an order of magnitude cheaper than McDonald's per calorie and they're non-perishable. Combine that with whatever fruit and veg is affordable fresh or frozen, a bit of cheap seasoning, and you're still coming out ahead.
You obviously need access to cooking and storage facilities to eat like this, but the target audience of McDonald's is the time-poor, the resource-constrained, or the depressed and disabled, rather than just the money-poor.
IMO, I think it breaks even, but eating out saves a lot of time! Healthier cooking at home? Yes. I studied this for myself (N=1), and my cooking is about US$10/meal give or take (asparagus, chicken, rice, water to drink). If you cook for two or more people, then I think cooking at home comes out ahead financially.
>cooking is about US$10/meal give or take (asparagus, chicken, rice, water to drink).
You must be eating an absolute TON to eat $10 worth of chicken, asparagus, and rice. I just checked the prices at Target and rice is $1.89 for 2 pounds, chicken thighs are $1.69 a pound. Asparagus is spendier at $5 for 1 pound.
How many pounds of chicken and asparagus are you eating? Even if you ate two pounds of chicken and the entire pound of asparagus you aren't hitting $10.
...and add the time for preparation, cleaning up etc.: Thats one of the most frustrating things when cooking for one person - you invest 45min to eat 5min and the rest is "organisation & logistics"
45 minutes is crazy. I have a chicken and rice dish I can make in 20 minutes (yes, I've timed myself because I'm weirdo and enjoy those chef shows). It takes 20 minutes because that's how long the rice takes. It can be faster if I use shrimp instead of chicken (more expensive though) and noodles instead of rice. It also makes ~3 servings.
Wow, i got downvoted for complaining about my cooking times on HackerNews, this is a real innovation:
so:
- 5 min walk to supermarket
- 10 min in there
- 5 min walk back home
- washing & cutting wedgetables 7 - 10 min
- maybe cutting some meat: 5 min on top
- eating 5 - 10 min
- cleaning up the kitchen 5 min
Haha...I didn't downvote you, but 45 minutes to cook seemed crazy to me unless you're making something new or complicated or you're socializing. You also kept eating separate from cooking in your first post.
I don't count the walk to the supermarket since even if you eat out almost every meal, you still tend to need to go to the grocery for items either way. And if you eat out, you still have to go there and back.
Cutting should be in parallel with cooking. Similar to cutting, cleaning should also be in parallel with cooking. For example, my knife and cutting board are washed, dried and put away before my sauce finishes simmering down.
That's why when it's just me I don't really do much cooking. I'll eat ultra-low prep stuff like toast (w/beans, hummus or avocado), bagged salad, frozen food, or grilled tofu.
I believe the long-term average in the US and UK was somewhere around 20 - 25 weeks so that's still broadly in line. Not trying to dismiss anyone but there is a cacophony of voices about the difficulty in finding jobs but hard to ascertain if that is any different from normal or we just got used to a boom cycle (ex Covid) and that's causing the disconnect?
reply