I have used a Windows OS almost every day of my life since 1999 or so. Last December I had a choice and switched to MacOS for work laptop. Since then I seldom use Windows and I don't really regret.
I still use an Xbox almost every day so there's that. In the last couple of weeks there's been some good news coming through for Xbox so we'll see.
Asha is doing some good decisions, or it seemed so until the XBOX Reset blog post.
However beyond XBOX, Microsoft is one of the biggest publishers as per amount of owned studios especially since ABK acquisition, so even when XBOX is doing bad, Microsoft Publishing is doing great.
For me, investing in hardware seems to be the way to go.
I learned coding nearly 24 years ago and still learning new stuff all the time. At no point in time I had to rely on a subscription model to learn and do new stuff.
If LLM and agents are the default tools for coding and building software, at least for next few years, it seems like a no-brainer to invest $2000-3000 on hardware, like a Halo Strix PC.
I wondered if there might be a no brainer "free" option on discarded hardware.
I have a GTX1080ti which i think is circa 2018, it's unused, more than paid for itself over the years, owes me nothing at this point so the hardware is free.
It runs Gemma e4b multimodal, qwen 3.5 8b or the qwen 4b embeddings models well enough (40+ t/s for the LLMs).
The machine consumes 350 watts at the wall when under load (3 watts when sleeping, 80w at idle). Electricity costs me £0.035GBP/kwh which is cheap for the UK (load shifting via house battery).
144k output tokens for around 1pence (and takes an hour to do that in theory).
It's only JUST cheaper to use than the far more capable deepseek v4 flash model despite the free hardware and ~10x cheaper than normal electricity.
Yes and no. Hardware does lock you in. Granted, I am happy with my 128gb of shared memory, but I am mildly concerned that it actually is more expensive now than when I bought mine. It does not bode well for the future; not when combined with recent WH admin moves on Anthropic and the reality that next batch of good models may require more than 128gb to run well.
edit: I am not dismissing local. I am one such user ( though I have subs too ), but one has to be clear eyed about the trade-offs.
$3k isn't getting you frontier model capability. It's barely getting you any capability if that's split into buying an entire PC rather than just GPUs.
With you here. I'm using my cheapo 16gig vram card I picked up a year or so ago, and I'm like -- yes, I percieve that you can pay for way more tokens per second that I can do at home.
But that feels like measuring productivity in lines of code. For what I'm doing, I'm not seeing the benefit in any subscription.
Sure, I can't one-prompt a whole new boring CRUD app, but oh well.
I hear this argument all the time. There's an excess of this, there's an excess of that. Seems it only comes from people who are not directly involved in hiring of such roles. We hired for an analyst role few months ago in the bay area and there was no qualified American applicants. My wife is in pavement consultancy and they hardly ever find qualified Americans for pavement design jobs.
People can’t seem to separate the issue of exploitation of H1B by (mostly Indian) consultancy mills to lower wages and bypass normal immigration vs the legitimate value of specialized skilled visas. You can fix one without killing the other.
It’s plausible Education might be one of the industries that gets exploited as they have no caps in the lottery like other H1 visas categories such as tech or doctors. But I don’t know enough about education visas personally.
Yep, H1B visa abuse exists and should be clamped down on, but it's also extremely vital to our sustained economic growth and frankly our biggest growth industry: tech
No, it’s clear that H1Bs have destroyed the American software industry as evidenced by the fact that the American software industry saw the largest growth in wages relative to any other field in the U.S. over the past 2 decades while also seeing a massive growth in the numbers of employees over the same period, also outpacing every other industry here as well.
Clearly the H1B has been devastating to American workers unlike all those other industries which haven’t seen an influx of a capped number of H1B workers every year.
And that’s even before we get to how the U.S. software industry remains one of a handful of industries where the U.S. is the leader in the world and has generated trillions in wealth, and is largely responsible for the US’s continuing dominance in the world today.
(A) individual must be interested in job/benefits/comp. and decide to apply. This makes them a ‘candidate’
(B) candidate must be qualified per minimum requirements
It’s entirely possible to not have candidates if no one is interested in the position at the stated comp/benefit rate.
But are there qualified Americans who could easily switch to that job for one reason or another? Is the issue a lack of qualified professionals or is it a lack of interest by qualified professionals in the listed position?
It's very easy to receive no applications from qualified professionals that do in fact exist by simply not offering to pay them enough (among many other things). That shouldn't mean you get to undercut the domestic labor market; rather you should be forced to rethink the business plan.
One of the problems is that geography and demographic movement trends within that geography is a very real thing.
Let’s say a rural town has lost population in the last 20 years, and most of the population that left is educated.
Now they need a teacher, which requires a bachelors or even masters degree.
The rural town’s unemployment rate is 10% but there are no qualified teachers who are unemployed.
So now we want to move someone in from a nearby urban center that has a big market of educated people. But the unemployment rate in that big urban area is 3%, and the area is wealthier with a higher salary rate for jobs across the income spectrum.
Let’s say my local teacher salary is $50,000, the big city teacher salary is $90,000, and the big city high school diploma career salary is $50,000.
I have to find someone who is a qualified teacher who isn’t already a teacher and isn’t already working someone where else that’s still paying better than my local area. Plus, that person has family, friends, and prefers the big city with all its amenities and infrastructure. I can tell you right now that you would have to pay me far above market rate to get me to move because I’m already employed and happy.
In contrast, someone in a foreign country is potentially getting a huge upgrade to move to the US or another developed wealthy country and is way more motivated to make that leap.
I imagine these programs exist because the cost benefit just makes sense. Not only do you solve the imbalance faster, easier, cheaper, but now the wider country has gained educated population which is generally an economic benefit.
Certainly there are flaws in the system that need to be fixed. I don’t mean to advocate for it necessarily, just explain why I think it exists.
I would also point out that it’s not necessarily the case that the local labor market is being undercut (see the geographic example I gave above), it’s being expanded, and that includes adding someone who is paying taxes, buying stuff from local businesses, etc, which they do even before they become citizens.
> I can tell you right now that you would have to pay me far above market rate to get me to move
If everyone feels that way, then it's not an above-market rate, it's by definition the market rate. The market rate for a job might be different at different locations.
(Note that I'm not taking a position on whether or not using the H1B program to reduce the market rate here is a good idea)
Yeah I agree with all that but notice that the comment I responded to was either about an analyst in the bay area or a stream of pavement design jobs in unspecified locations. I wouldn't necessarily object to the metric of "do qualified americans exist" being limited to a certain geographic area as long as the resulting criteria was sufficiently difficult to abuse.
The bay area's existing talent pool is hugely immigrant in nature. It's totally within expectations to put a job post for an analyst of some kind there and only get immigrant visa holders applying. They also make a lot more than many Americans who wouldn't be qualified but still live in the area.
Isn’t it a self-fulfilling issue? Dependence on H1B and other visa dependent workers leads to lower salaries which discourages local talent from that specialty.
What requirements did the role have and what’s the salary range?
From what little I gather from online job listings, most foreign labor dependent positions are trying to pay 90K for a masters degree, maybe 120-140K Bay Area. Additionally, many of these job listings want extremely specific degrees or certifications that frankly are of little interest to US citizens - but F1 students will take any masters program despite the program having little salary benefit - the degree is a requirement for the visa.
I have a hard time believing you can’t find a US civil engineer who could learn the subject matter right out of college. Although saying that I know first hand low starting salaries have pushed students towards mechanical engineering or CS if inclined.
> What requirements did the role have and what’s the salary range?
4+ years in product development.
Python/R + a low-level language.
Terabyte-scale data stream and batch processing.
HPC knowledge (vectorization, memory access, distributed computing) to build efficient algorithms.
Degree in a quantitative field (Math, Stats, Physics, CS, or Engineering).
Upper limit on compensation was 200k.
> Although saying that I know first hand low starting salaries have pushed students towards mechanical engineering or CS if inclined.
You answered your own question. The American engineering pool consists mostly of high school diplomas who can't pass PE exam at multiple trials.
Edit: coincidentally, my wife was offered a state civil engineering job in Bay area. Didn't take up because the salary offered was below 100k, even with 5 years of experience.
There are a ton of jobs that pay as well or better with lower requirements, even outside the bay area. Anyone with that level of experience, Python and a low level language isn't going to take you up.
I'm not normally in agreement with the "you're not paying enough, there's plenty of people" crowd, because I've been on the hiring side too and know what a crapshoot it can be... But you're definitely offering too little for those requirements.
I responded to the point someone made- there's an excess of workers in America. Firstly, when there's an excess, wages are supposed to lower, even for Americans. Secondly, even if there is an excess, there was no evidence of that in my experience. In addition to the full time role we also interviewed interns in fall, and in my experience they were all either immigrant or children of immigrants.
You don’t see an excess of workers because the compensation was too low. Your requirements were such that you were realistically competing with Meta and Google offers for $400k+ and $200k was your max possible compensation.
I did not post the full job description word for word to not doxx myself. But the job description explicitly mentioned that anyone with the right aptitude should apply, don't have to qualify for every requirement. We hired a recent graduate with relevant research experience.
$200k is very middle of the pack for a salary offer in the bay area, and most places will push total comp up with stock and such, whereas OP mentioned that 200 was the UPPER bound, meaning they wouldn't be offering it to a junior developer.
I may have also misread the degree requirement as being higher than it was, but I think my point (prior to the edit) stands- for the posted requirements, the offered salary is low compared to other available jobs.
I'm sorry you all have lost your minds. I live and work in the bay area. $200k for someone who is 26 is far more than enough and should absolutely be able to get you a qualified person for the job description given above.
It's a middle of the pack/low salary offer in FAANG but the vast minority of developers here work at a company like that. It's hard to remember that sometimes.
I mean I don’t have a horse in this race, but I don’t think this is a good example.
If this is a senior enough position to justify expecting this level of specialization, that compensation is not nearly high enough, so issuing this H-1B would add downwards pressure on the compensation of American worker.
If this is not a very senior role, the American worker’s interest is that you find someone with a less specific background, compatible enough so that they can be trained.
> Dependence on H1B and other visa dependent workers leads to lower salaries
It does not generally. H1B employees are more expensive usually. In most companies - like any notable tech company - they are paid exactly the same due to fixed compensation plans, but cost more to the company once you include legal fees, processing fees, and especially the time delays and risks. It’s not even close in terms of a cost comparison. This isn’t a controversy among people who are actually involved in hiring and compensation - it’s well known. But this perception persists.
They don't have to undercut their coworkers on an individual level. When a position that could otherwise reasonably be filled by an American is gated for any unnecessary reason, be it undesirable pay or excessively specific requirements or whatever else, that effectively removes that position from the domestic job market. When that is repeated many times the end result is the same number of domestic applicants competing for fewer positions. That results in downward pressure due to basic supply and demand.
That's fine if there's a genuine need for a specific sort of specialist and the US simply isn't producing enough of them. But when it's silly hyper specific requirements it becomes detrimental.
tech salaries have only increased and never decreased. Not in 2000, not in 2008, not in 2020.
Your narrative runs against the facts on the ground.
Not only foreign workers were part of the reason US tech dominates the world, they also greatly contributed to every single major invention in tech, including the Attention paper which led to Transformer based LLMs.
All of these stupid anti-immigrant narratives are exploiting either ignorance (foreigners undercut wages, while wages have tripled and more) or motte-and-bailey tactics (WITCH bad and H1B fraud bad, lets abolish all immigration)
That's highly deceptive. To start with, salaries have to increase just to tread water given inflation. Beyond that, there's nothing wrong with a position that produces a large amount of value being compensated in accordance. If a particular sector does well we should expect salaries to rise relative to the economy as a whole.
The question is simply whether a policy would depress wages relative to not having that policy, and it's merely one of many factors that should be carefully considered when weighing the costs and benefits.
> All of these stupid anti-immigrant narratives
You reveal your own biases. Personally I'm originally from academia. I'm accustomed to a workplace where citizens are only barely a majority and of those many are naturalized immigrants. I have no qualms with importing labor that's significantly more skilled than the average american in cases where doing so benefits our society on the whole. Neither do I have qualms with filling jobs that are legitimately unwanted or that we truly don't have sufficient local talent to support. However I'll note that the last one there is exceedingly rare.
not having H1B foreigners like Elon Musk or Satya Nadella create/grow companies that employed tens of thousands Americans and created trillions of value for investors.
Remember it was White American Steven Ballmer who almost ruined Microsoft (he took MSFT at $620 bln and left at $280 bln valuation), and it took a foreigner like Satya to bring it to $4 trillion.
not having foreigners who could discover Attention mechanism or Transformer based models that led to AI boom.
US would literally be indistiguishable from Europe's lackluster development and growth, would you want US look more like EU ?
All of the wignats who complain and bitch about immigration completely forget or take for granted all the tech progress made in the US by foreign workers, and all the non-tech jobs that get created thanks to tech work being done by foreigners.
Just visit any graduate program in the top tier US university lab, visit any leading scientific conference, or lookup nationalities of top cited research papers, and imagine USA without all those talented people with foreign sounding last names.
quite the opposite, all of H1B opponents imagine current wealth and state of tech will preserve the moment immigration stops and their loud temper tantrums are satisfied.
that American capitalists are stupid enough to overpay underskilled Americans, the moment H1Bs are banned or limited, instead of opening satellite offices outside US
Elon Musk is causing way more harm then benefit to America itself and its economy. I dont think he is a good example of a positive effect of immigration.
"We were unable to attract highly experienced data analysts, and we REFUSE to ever train anyone for any role, so we should be allowed to use scab labor to undercut American wages"
No such issues. It's actually not a solo business, it's a civil/geological engineering consultancy firm with a mix of state/local government and private clients.
And in China they were far more occupied with a cultural revolution than any form of advanced education. At best you could argue that they were teaching basic literacy to more people.
Tough. The undesirable consequences needs to be forced to occur so that the voters have no choice but to deal with it. You shouldn't get to just rip the bottom out of the market and proclaim to have fixed the problem.
This is pretty much how I've been operating. While the C-suits have been always encouraging everyone- technical and non-technical folks alike- to use AI, the ask from my manager and skip level has always been for deterministic output. Before last December or January I was mainly using LLMs for autocomplete, whereas now it looks more like "given this input write script to generate this output", and after some corrections, "summarize/update this session into a skill". Script for future humans, skill for future agents.
AI is absolutely directly responsible. Managers are literally asking whether people need AI credits or interns over summer. Most people are taking credits, and internships are getting cancelled left and right.
> Managers are literally asking whether people need AI credits or interns over summer
That's a bit like asking them to choose between AI credits and a visit to the dentist. No engineer has ever hosted an intern in the hopes of improving their project's velocity; people host interns because "it's the right thing to do" for the industry, because it's psychologically satisfying for many people, and to build cred and/or experience as a mentor. And companies nudge their employees to host interns to hopefully influence potentially-valuable-in-the-future smart youngsters to come back as a year or two later as full time employees.
Where I work, interns are hired to work on projects that a senior person could do and failure won't wreck us. And your last sentence is where we see the real payoff: the bright ones come back and become valued team members.
> people host interns because "it's the right thing to do" for the industry
This is bullshit, assuming that we are talking about for-profit corporations.
In my experience, the reason why my teams have hired interns is to get a solid, multi-week preview of their potential quality and abilities. During the hiring phase for juniors, the signal-to-noise ratio isn't very good, but a good intern can get a lot done in one summer. You can easily pick the best one or two interns to hire when they graduate. Then you dramatically reduced junior engineer hiring risk. Also, if they suck or are disappointing, then the loss is minimal -- don't hire them. If they are really awful, then just throw shit work at them and don't waste time trying to mentor them. Really, it goes both ways.
> And companies nudge their employees to host interns to hopefully influence potentially-valuable-in-the-future smart youngsters to come back as a year or two later as full time employees.
And there's less incentive to do this when you anticipate needing fewer employees.
His point is that the engineers wanting to opt for "not intern" isn't really a data point on whether interns are helpful. It may instead be a data point on the propensity for people to opt out of work when they have a good excuse.
Wouldn't it go the other way? Instead of working, you're "mentoring the intern" over a long lunch and telling them long meandering stories about company lore.
In my experience hosting an intern does not count much towards your review. So sure, but you also could have done nothing and your review would probably be the same as if you hosted an intern.
It’s possible somewhere does properly incentivize this, but the companies who were regarded as doing it the best still don’t in my experience (at least once the company gets bigger than a startup).
> Managers are literally asking whether people need AI credits or interns over summer
My Brandolini's Law alarm is ringing loud and clear here. I would say the answer is both -- hire the intern and give them a bunch of AI credits. Ask them to work through the old backlog of shitty, low risk work that no one else wants to do.
This opinion should only come from a CS grad who is accustomed to building CRUD apps and allergic to any real world problem domain.
At my work in insurance solutions we're solving perabyte scale problems. We do not have a Rust developer in the team, but taught ourselves with the AI and now automating workflows which took days to under a minute now.
If you look at the speed a company rolls out new products/features, it should be obvious that "interesting ideas" especially good ones are infrequent and rare.
No offense, "using Rust to address a performance problem" is by no means novel or interesting. That is just another person's CRUD.
Using Rust is actually the most boring part of the problem. It would be almost impossible to solve for anyone without certain level of domain expertise.
I have limited enterprise budget and Claude 4.7 costs 7x more. So unless there's close to 7x improvement, it doesn't make sense to switch to 4.7.
I actually gave both 4.6 a really complex task. It kept on thinking for several minutes before I hit the brakes. I then gave 4.7 the same task, and didn't notice any difference in behavior. Clearly not worth the 7x premium.
I hope 4.6 becomes cheaper/free at some point because I'm starting to see a push towards optimizing token expenditures across the board. While frontier models are still the default for developing new workflows, everybody is starting to ask how to automate repetitive tasks without using tokens.
Our household income is 300k, and I took a big risk purchasing a home in Socal where mortgage is 50% of our take home income. In a space of 1 year or so, we went from saving two out of four paychecks each month, to sinking two into mortgage each month.
For me, we have one kid and we plan to stay put for at least 10 years. It's a good school district. The quality of life is excellent considering the weather, outdoors, cultural diversity, things to do, and proximity to international airport. We have friends and family here.
But to me what mattered most is that I am 40, our income is going to plateau. Renting is good advice, but in 5-10 years, we won't be able to afford rent here, let alone buy. On the other hand, I can refinance now and bring my mortgage down to what I was paying for rent previously, and in 10-15 years be mortgage free at a place with all the benefits I mentioned above.
Maintaining your own home is a great hobby for guys 40 years old and older. The house may age, but your engagement and knowledge of your house steadily grows. Even contracting out the work has its interesting elements. Keeps you active as you age, both mentally and physically.
The key for older homeowners is finding your querencia, the place of your heart. If where you live is your querencia, then countless options open up for you. Simply living in your home and community becomes a joy. Your community becomes an extension of your home and your heart and you become a vital part of that community. In your golden years, that community provides deep meaning and grounding when your work life quickly fades. Building that connection takes time and energy…plopping yourself somewhere else in your 60s, you may not find it. Building it now and in your 50s guarantees it’s there when you want it.
> Renting is good advice, but in 5-10 years, we won't be able to afford rent here, let alone buy.
You don't know if you will get priced out. You don't know how housing prices will evolve in your area. They could as well go flat or fall. That uncertainty about local affordability can suffice as an argument for buying. Just don't confuse it with certainty about your future ability to afford the local area.
I still use an Xbox almost every day so there's that. In the last couple of weeks there's been some good news coming through for Xbox so we'll see.
reply