The reality is that most people are paid for their time, not for their output. I think most contracts for salaried employees are along the lines of "work n hours a week". If you want to get paid for output, you can't be a salaried employee.
Salaried employees aren't paid for their time, they're paid for a combination of their output and their availability. Availability used to be strongly coupled with time but technology has introduced some flexibility there.
Correction: Their PERCEIVED output, not their output. Output can usually not be objectively measured in IT.
Storypoints? Overestimate wildly. Stories done? Split stories like atoms! Impact of work done? Present your button color change stories as company saving divine intervention!
Can you give some examples of jobs that are usually done by salaried employees and are paid by output? All the examples I can think of, are usually done by independent contractors.
Not by output, for output - as in, not paid per unit of output, but paid for achieving a sufficient level of it, the definition of which frequently changes. That's baked into the definition of salaried.
I don't think this is the reality. It is part of it, but people get paid different salaries, why? Some are more productive than others. Aside from leadership's (and society's) biased ability to determine value, these people theoretically get more because they contribute more.
This is a naive view. Companies will literally pay you as little as they can get away with. If they think they can hire someone “equivalent” according to HR for less money they will often do so. Actual contribution means nothing unless you’re willing to walk away. I’ve literally seen an intern with less than one year of experience when hired into a company where a senior architect who has been at the same firm for 16 years made over $40k less than them. The former intern was great. I had no real complaints. And he was more ambitious than the senior architect. But he stayed with the company two more years before jumping ship for another significant pay raise while the senior architect is still there making about the same amount as he did five years ago. It doesn’t matter how much value the senior architect delivers if he doesn’t demand his increase in pay.
> Companies will literally pay you as little as they can get away with.
Everyone pays as little for everything as they can get away with.
Companies aren't doing anything unusual or evil. This is literally just how a market works.
A lot of people forget that the job market is a market. Or they wish it wasn't a market and they could block out competition for themselves so companies had to pay them more (at the expense of those excluded)
This might be fair if the rules of the market were static, but companies are actively using their power and influence to affect the market rules so they can get away with paying workers less and less - that's not to mention collusion and corruption.
Just because this has been normalized doesn't mean it's not evil. If a healthcare company introduces processes that deny people life-saving interventions just to funnel more money to their shareholders, that company and its leadership are thoroughly evil.
I'd also like to add that I often pay more for things than I could get away with. I try to prefer locally owned businesses when possible, and if they have fair prices and give me good advice I often pay a bit extra (essentially a "tip" for the business) to support them.
> but companies are actively using their power and influence to affect the market rules so they can get away with paying workers less and less - that's not to mention collusion and corruption.
What do you mean? Labor is a market. Companies can’t force anyone to work for them. If the pay gets too low, people leave for other companies or jobs.
Just like all markets, the labor market isn't an immutable natural structure. It is actively shaped by the laws and regulations of the market's respective physical location. There are too many specific examples to count, but a few of the most obvious ones are IMO: Child labor, minimum wage, max. working hours and/or overtime, worker safety & health, ...
> Companies can’t force anyone to work for them. If the pay gets too low, people leave for other companies or jobs.
In an ideal world, sure! But in the real world, there are many sources of friction that - when added together - keep people tied to a job, even if they feel the pay is too low.
Markets are social constructs, not something given in nature. As such, there are rules enforced by laws to how the markets operate. Powerful money interests can shape those rules by influencing politicians to pass laws in their favor.
> Everyone pays as little for everything as they can get away with.
Bullshit cop out. I regularly spend more than the “average” on technology because I’ve been burned by the bottom of the barrel too many times. Maybe you cheap out on all of your purchases, but it says nothing about “everyone”. The fact that you claim “everyone” has as few moral qualms as you is telling. But not in a way that reflects everyone. Just short sighted selfish assholes.
> Companies aren't doing anything unusual or evil.
You literally just described an evil behavior. It's obviously not evil when you're shopping for gadgets, but if you think that's morally comparable to the livelihoods of human beings, I'm sorry to inform you that you're a psychopath.
> This is literally just how a market works. A lot of people forget that the job market is a market.
You say that like you think it's a law of physics. Markets don't exist in nature. Markets are a conscious, deliberate creation by humans.
> Or they wish it wasn't a market
Correct! We've unambiguously determined that applying market logic to people's livelihoods is inhumane and immoral. There shouldn't be a market for being allowed to live. It should just be guaranteed to everybody.
In my experience, productivity effects on salary is a rounding error. There are only two significant contributing factors for salary: 1. Your home address and 2. Your visa status
Productivity might get you a 5k raise more over your colleague on a 160k salary. Meanwhile the same engineer in Taiwan is more productive than you and getting paid 40k while working for the same company on the same product, and putting in more hours to boot.
Or the engineer could be working for a certain company in South Korea and get a bonus multiple times their annual salary because of their union contracts tied to profitability.
while productivity is correlated with salary, generally the ability to ask for a raise, to defend your pay and office politics navigation would be more impactful on average
I feel a bit like some have misconstrued my comment. I reread and can see how it could be read as quality of work being ALL that matters. Indeed, salaries capture many things and many bullshit ones. I don't exactly think what I wrote precludes that racism, sexism, nepotism, or cozying up to your boss factor more than they should, but I can see it. If I had another shot, maybe I'd reframe the provocation as "by and large Senior Xs make more than Junior Xs within the same company within the same geo, why?" Perhaps the existence of senior/junior titling at all is based on some sense of output/pay correlation.
They are paid for what they collectively output. The only reason that people seem to be paid by their time is that it's hard to measure each one's output individually and granularly.
Exactly. Hence why productivity has increased while wages haven't. Wages have increased alongside the productivity increases of those wages, of course, but wages are decidedly for human labor, by very definition, not robot (computing, automation, whatever you want to call it) labor. If you want to capture the value of robots you need to use your human capacity to deploy robots, not to sell your labor.
That sounds like a roundabout way of saying that the capital owners have captured the gains from productivity without increasing the compensation to the workers.
But this belies the fact that the workers had to grow more skilled to operate and maintain those machines. They took on additional costs in education that are not being compensated. They're being asked to get more work done by being higher skilled, but the bosses are collecting all of the additional revenue generated.
> That sounds like a roundabout way of saying that the capital owners have captured the gains from productivity without increasing the compensation to the workers.
What do you mean in a roundabout way? It says that quite explicitly.
> But this belies the fact that the workers had to grow more skilled to operate and maintain those machines.
No they didn't. What is a capital owner going to do if the labor didn't become more skilled? Just shut everything down and take the ball home? Ha ha. No. Not a chance in hell. This is the silliest thing I've read all day.
> They took on additional costs in education that are not being compensated.
One wonders why. When they said "go to college to make more money" they were clearly talking about using college research labs to develop capital. Exactly so that you become the one deploying robots, not selling your labor. That is where the "more money" comes from. How people — traditionally smart people, even — got it into their head that meant go to college to get a low paying job, the same job they would have gotten anyway, boggles the mind. I suppose smarts don't transcend into every subject.
There is a lot of regulatory stuff, particularly around benefits, that push people towards nominally 40hr salaried contracts even if they don't need all 40 of those.
"Salaried" vs hourly is increasingly a scam anyway, but all that benefits stuff is something that would have to evolve. And it could, if people find the political will.
What regulations prevent an employer to give benefits to employees before 40 hours? And how do some employers give benefits before 40 hours despite them?
The contracts I've seen have an explicit floor, not a de facto one. I.E. The contract says the minimum number of hours you need to work. Some countries also have overtime laws which create a ceiling.
Either way it doesn't change that being paid for your output is the realm of entrepreneurship and submitting bids for project work.
Employment contracts are virtually nonexistent here in the U.S. While the norm is “40 hours”, actual requirements vary by employer, may or may not be communicated in writing, and can change at the employer’s whim. If you make under $107,432 per year (which is around starting salary for fresh graduates at FAANG), you are entitled to overtime pay (minimum 1.5x) for hours worked beyond 40 in a week.
At least for my software job in the US, and other salaried jobs I’ve seen, there are explicitly no hours listed, and it’s supposedly based only on your output. In practice though, if your butt isn’t in the seat 40 hours a week or so, and usually more, the boss will be mad.
At least in the US, there are regulations around what constitutes "full time" employment, and many benefits (such as insurance) are tied into being offered only for full time employees, or at different tiers between part and full time.
As such, you are still expected to work a minimum amount of time. That's what you're signing up for. Fixed deliverable contracts- completing certain objectives- tend to either specify those things as minimum performance expectations, or are for contractors rather than employees.
At my most recent role we were definitely being judged by output metrics (both jira tickets completed and github prs merged). They showed us the jellyfish tool they used to check those metrics. Well, some of it. Regular ICs themselves didn't have that access.
When you work for a startup, or a zero-to-one project, it's hard to say you're even paid for your output. You're paid for probabilistic/expected future value output.
Every meeting, every memo, and every prototype is output in terms of the employees doing that work. Whether it's directly saleable is irrelevant. The investors base the value of their investment on the expected future value of the company, but the people being to do the work are being paid for the work they are doing regardless of what the future value of the company becomes. That is if they are paid a salary. If they are given shares, then that compensation is entirely dependent on future value.
You can demand whatever you want. You could demand a million dollar salary if you wanted.
The challenge is that there are a lot of very qualified devs who would do it for less.
Labor is a market. Supply and demand determines your wages.
There are always hand-wavey arguments about unionization fixing this, but when other developers are hungry for those jobs and willing to go around the union to work them for pay then that doesn’t really work at scale.
There are several unionized software development groups in the US. They don’t have a good track record of getting significantly higher pay or even getting their demands met from their limited strikes.
I was in the musician's union for 12 years before I got into tech. There were some silly rules, like someone couldn't be both a musician and and orchestrator on the same show, because it's "doing 2 jobs". It's like saying you can't be full stack. You couldn't fire people who were bad at their jobs and stopped putting in any effort. There was a profit sharing agreement that the union rejected, because it would come at the expense of higher base salaries, and then they wondered why there were only big producers that first developed the show out of town.
Some rules I actually liked. Rehearsals started and ended _exactly_ on time to avoid overtime (showing up late was the only reason you could be fired, which was a useful compromise). But generally, the union was the yin to the producers yang, and an adversarial position as worker advocate was where they wanted to be, they didn't want more ownership.
If someone gave me the chance to join something more like a worker-owned coop, where the workers on the business and vote on how it works, I would actually be down. There's a grocery store down my street like this and it's a great place. I don't know how this would actually work in tech. If there's no startup capital, no one will have a salary or benefits for years until there's a profit (if at all). And capital comes in exchange for ownership of the future upside.
> If there's no startup capital, no one will have a salary or benefits for years until there's a profit (if at all). And capital comes in exchange for ownership of the future upside.
The developer co-op projects I've seen have targeted consulting for this reason. The idea is that developers get together and start doing projects that can bill clients immediately, and then they'll pool the money back into developing their own something later.
In practice there's no real difference between a group of people consulting together and a co-op of consultants when everyone is just billing hourly at the start. Nobody really wants to spread their earnings around the co-op because you can see the relationship between hours worked, hours billed, and dollars coming in so clearly.
> and then they wondered why there were only big producers that first developed the show out of town.
Most unions derive a lot of their negotiating power from location-based constraints. You can gather enough musicians in one place to form a union because there are a limited number of musicians within driving distance of the location. Musicians can't do their performance over zoom and the job can't be outsourced to another country.
Software jobs have no restrictions like this. Every time there are calls for unionizing software devs, nobody wants to answer the hard questions like what incentives multi-national corporations will have to cater to the unionized employees in a country like the US where we're already paid more than our international counterparts. It's just assumed that the union will form, then companies will have no choice but to accept their demands.
> If someone gave me the chance to join something more like a worker-owned coop, where the workers on the business and vote on how it works
I run a co-op, and the most surprising thing I've learned is that it seems the only reason other people don't set up their businesses as co-ops is, frankly, greed.
In this startup ecosystem, why would you do a co-op when you could instead chase that multi million dollar exit that profits you? If we have any kind of event like that the earnings would be distributed basically equally among around 30 people. 20 mill payout suddenly becomes... Less...
Anyway though apparently statistics are on my side. Apparently co-ops are more sustainable and live for longer than traditional corporations. It seems to me that people are much happier here than at traditional corporations.
Clients are happy, engineers are happy and productive, I'm happy because I don't have the entire responsibility of the business on my shoulders - engineers are constantly contributing to the improvement of internal systems for example, because why wouldn't they? That saves them money too!
Downsides are increased ownership come with increased responsibility. Can't just clock in here, people need to manage the client relationship, issue invoices, participate in the accounting, and if they want more work when their gig is closing out, either try to sell the client on more work or help us find another one. Until we can find an alternative revenue stream to selling our labor to clients, we're all beholden to keeping the BD wheel turning in order to get paid. Upside is that we're keeping 85% of the margin for ourselves rather than at toptal where you keep, idk, 30%? And that 15% is still our money anyway it just gets used for overall co-op stuff, which members get to decide on.
Also I don't know if this kind of business lends itself to the silicon valley mythos of the huge work week for a few years followed by functional retirement. Instead it seems we'll all need to keep working for the foreseeable future, but at our own reasonable pace, fully remote, our own hours, at very high compensation but not "fuck you money" compensation. Well except for our people in Ethiopia and Taiwan, for the local market rates, their compensation is getting way up into that territory.
That said, I don't know about our ability to survive a full on capitalist attack in the form of lawfare or getting priced out or closed out of deals. Priced out in a labor market would be difficult since the only way to beat our labor margins is to hire directly, but we could be lawfared to death pretty easily considering our ARR might be 1 mil this year if we're lucky.
Maybe i’m a cynical capitalist, but I expect everyone to be operating in their own self interest. There are few true saints. Startups are high risk high reward. 9/10 fail. If a VC gets less upside when it works, the successes wouldn’t pay for the failures.
Co-op for consulting actually seems like it could work.
> Co-op for consulting actually seems like it could work.
I made a comment about exactly this in the thread. This is where a lot of co-op ideas go, but they run into friction when they get to the part about sharing the money around. Consulting is one place where developers can see a direct relationship between their hours worked, hours billed, dollars coming into the business, and dollars going into their paycheck. There are some ideas like putting shared money into a group fund to cover vacations and similar expenses, but generally a consulting group is money-in money-out. Nobody really wants to join a co-op and see the hours they bill go to pay someone who works fewer hours, so the money relationship gets even more tightly coupled than at a traditional salaried company.
> Maybe i’m a cynical capitalist, but I expect everyone to be operating in their own self interest.
Down to explore this with me?
I often wonder if this is self fulfilling. I'm not saying you're selfish, I mean, has capitalism trained people that all people are inherently selfish, and then, within this system where selfishness is rewarded, incentivized further selfishness and continued to sell this myth?
How much maintenance does capitalism take? If human nature is inherently selfish, and capitalism is a force of nature, a ground truth, the answer should basically be no maintenance, right? Is that reality though? It seems to me it takes a lot of maintenance, the myths of capitalism. Look at the incredible amount of resources the USA dedicated to fighting Communism, both explicitly (crossing swords with the Soviet Union) and in the background (overthrowing socialist leaders in south America, meddling in southeast Asia).
Is the key difference for me that led to me making a co-op (rather than a traditional geographic arbitrage agency) that I don't buy into the capitalist notion that humans are inherently selfish? Maybe I just happened to have the experiences that indicated to me that this isn't true, that humans are inherently mostly selfless and social, which led to me researching this and learning that history supports this notion of humans as social-first organisms.
When I tell people about my business, a lot of questions I get confuse me: "You don't require timesheets tracking granular work? How do you know people are actually accurately recording the hours and not 'stealing time?'" the answers are always basically, "because I trust my co-op members," and "I really don't care if people are 'stealing' from me." Capitalism teaches us to always "get ours," but as soon as you let go of that, if people trust me as a genuine actor, they seem to abandon these unnatural principles and work with me honestly, as humans naturally want to.
For your VC example, I mean, nobody will invest in us. Maybe that's the double edged sword - you stop playing within the rules of the game, people won't throw the ball your way anymore. There's no massive value to extract here. That probably limits our ability to rapidly develop some products to generate revenue outside of selling engineering hours. Maybe that's ok though, the business might grow more slowly but sustainably? And maybe that cements our reputation as "no really, actually a co-op that genuinely doesn't put profit first?" I don't know. We'll find out.
But for you, what do you think would happen if you started putting yourself forward as a non-participant, non-self-interested? The prisoner's dilemma equivalent of saying "I will always cooperate no matter what you choose," and then following through on that? Is there a sort of aversion to that knowing that someone might exploit this aspect of you at some point, where you might end up in a situation where your personal profit wasn't maximized?
I see this as a bunch of unsolved game theory problems. USA vs Soviet Union I just see as a battle of competing empires. There have been competing empires for all of history, we haven’t figure out how not to do that yet.
At the smaller scale, I’ve lived in coliving housing most of my adult life. Mostly positive, but problems that make people retreat away from cooperation. Once a week potluck but otherwise people cook their own food. Chronic mess/clutter/dishes/laundry headaches.
But capitalism (and communism) are relatively recent economic developments. Most of human history was some form of hunter-gathering where cooperation inside the group was necessary for survival, with varying levels of cooperation and competition outside.
Should note that both the West & East included a lot of economic cooperation among countries within those competing Cold War economic models, even if there was coercion involved (and not just from the Soviets).
My personal mental model is that there is a quite bimodal distribution of people in the world. Many seem to see it as a zero-sum game, some others not. As father of a disabled child I come into contact a lot with the latter and can confidently say that there is a LOT to gain once you stop starting to stare at money only. Once your needs are basically covered, giving can give you so much.
React does have its benefits, but there is also a tendency to pick it because of the inertia it has over whether it is the best tool for the job. "Everyone uses react and this will maximize our hiring pool / set of contractors we can use", "A react project will look good on my resume".
I don't see why "inertia" or adoption should be excluded when deciding what library to use for the job. If you are manufacturing a physical product you will need to consider how easy it is to source any component before adding it to your bill of materials. I don't see why software engineering should be different. There is a reason why we are all working in boring languages and not Haskell or LISP despite their supposed technical superiority.
Openrouter is great for experimenting with models. I did exactly what you're saying to test smaller models that will run on commodity hardware and determine if it might be worth it to drop $10k on hardware. For me the answer was no, but it's close. I'm very excited for the next few innovation cycles to arrive.
UK sovereign data? Land of arrests for posts on social media? Member of five eyes, "you spy on our citizens and we'll spy on yours and call it intelligence sharing"? Land of the infamous Online Safety act? That UK? Why would anyone want their data in the UK?
Having data processed in the UK is not for privacy reasons, but legal reasons when working in Govt. fields such as legal/healthcare/security. For privacy most companies choose juristications with better privacy laws such as switzerland.
There is a long list of countries other than the US and the UK. I will go to bat for the US on this one though and say that one might want their data in the US because of the first amendment. Even for people that reside in the UK, what is the selling point of having data with them in the same country?
For all its flaws and limitations in practice, you have far greater data privacy in the UK than the US. Largely because it seems you basically have none as a private citizen in the US.
Just to pick on one thing, do you think incitement to murder people by setting fire to hotels, acts that subsequently happened, should just be shrugged at?
What I think about the particular case you're talking about is irrelevant. I guess we'll see if allowing the government to police speech is still such a great idea when Reform and all future governments are in power.
It seems that del monte proper is not actually declaring bankruptcy, so how is it that the American tax payer is left picking up the check on this one? Privatized profits, socialized losses!
Additionally tier 1 networks legally not allowed to route packets that don't include an attestation that they were produced by a computer running only approved software.
I said speed was great, Cerebas and Groq can provide better performance, likewise Fast versions of Cursor's Composer and Claude.
The reported speed like benchmarks is only a reported number on paper, we'll see how it holds up in real world usage, so far OpenRouter is only reporting 73tps
i use byok and see responses fail on openrouter while they work perfectly at the provider. the provider is often listed as 'down' and it's very clearly up on the original api and serving requests.
cerebras quotes oss 120b at 3000tps and it is under 800 on openrouter.
same with fireworks, i am getting much higher numbers not on openrouter. but recently i think fireworks deepseek is kind of spotty, the main provider i know that just doesn't go down is vertex and they charge 2-3x the rest
If someone from Grok is reading, don't waste time on these chaff features. The market will eventually deliver better 3rd party solutions to all of these things. There is an audience that isn't interested in these walled garden features and are only interested on intelligence per dollar.
Lol I wonder when Anthropic discussed the idea of Claude Code internally, were there bozos saying "3rd parties will eventually deliver this so we shouldn't waste time one it."
Personally, my work doesn’t want to get locked into a single LLM provider so we use Cursor. Much easier to fight the big corp software approval battle once then switch around the LLMs to the new hotness (provided legal has the requisite data sharing agreements in place, we’re not supposed to use Chinese models or Grok) but I can switch between Anthropic and OpenAI models at will.
Power users are hotswapping these models into their own agents (hermes, openclaw, etc) which have their own systems for project management, memory, interacting with tools, etc. The important metric is intelligence per dollar. Can I drop this model into my harness and have it be cheaper without losing intelligence. That is where the puck is heading.
What are good harnesses? I haven't yet been able to get good agent teaming approaches out of other harnesses yet, before that feature I mostly regarded the space as competitive, but until another harness can do as well with Claude models it seems like it's better for now?
Aren't they 'wasting' time on these features exactly because the engineering requires a different, more traditional skillset from the ML work model people do, and can be done in parallel?
reply