A lot of Japanese space companies have come under scrutiny for discrimination against non-Japanese, sexual harassment, and gender discrimination, often by the ex-NASA employees they hired.
A common practice is to train a transformer model to control a given robot model in simulation by first teleoperating the simulated model with some controller (keyboard, joystick, etc.) to complete the task and create a dataset, and then setting up the simulator to permute the environment variables such as frictions, textures, etc (domain randomization) and run many epochs at faster than real time until a final policy converges. If the right things were randomized and your demonstration examples provided enough variation of information, it should generalize well to the actual hardware.
Newbie here; do you mean they keep raising, so it looks like an IPO will never happen, or do you mean having a series F round will inherently destroy the value of the options they have given?
You flip a house, but instead of turning it around in a few months, you want to make it the most expensive house in your town.
You convince investors to front the money. You progress well and quickly. You could flip it now for 10% return. Not a massive return, but it’s quick. Investors are also concerned that if they don’t give you the money now, you’ll just walk away and they’ll loose their entire investment. They give you more money.
This process repeats several times. Until you’ve 10x the value of the house. The shit hole you bought for $100k can now sell for a cool $1M.
Massive returns, right! Well, not exactly. Turns out you spent $900k in the process. Your investors still want their 10% return. Instead of a $10k profit, they’ll looking at a $90k profit. Bigger pie for them, but that doesn’t leave much for you.
I'm in my 30's now. I did the same exact thing, with the same completely whacked out sleep schedule, when I was transitioning 18-19. I spent thousands of hours on Oblivion. My parents did not get on my case about it. They did occasionally say "don't be late for X," or "you're going to miss Y," but they never blew it out of proportion to what it actually was, which was "just a phase" for me. I had very bad grades at this time, and gaming was a way to cope an escape; not the cause of it. My advice is to give him some space and let it play out. If he's going to fail something, let him fail and learn. Let him have a bad semester. He will bounce back. Perhaps this is unavoidable at this point.
Reminds me of the cause of the Apollo 13 oxygen tank explosions:
"The high temperature emptied the tank, but also resulted in serious damage to the teflon insulation on the electrical wires to the power fans within the tank. The exposed fan wires shorted and the teflon insulation caught fire in the pure oxygen environment."
“Mostaque had embezzled funds from Stability AI to pay the rent for his family's lavish London apartment” and that Hodes learned that he “had a long history of cheating investors in prior ventures in which he was involved”
It also seems like the company just isn't doing very well, it looks like there have been constant problems since Stable Diffusion was released - not enough funding, people leaving, etc. Which I don't get - you create a massive new piece of software that is a huge leap forward and you can't just get more funding? There have to be big structural issues at Stability.
Yet for the 6 months before that they were talking about running out of money, and even the month after they got funded were considering selling the company due to money issues.
I’m practicality shaking my head in disbelief at all the red flags this guy has and people are still defending him. Stability and its work are great. We should support an open community and ethos. And Emad can still be a shady narcissist con man. These are all compatible views.
You are lucky, I got nothing for heavily suggesting a ~1M/year saving.
The only "personal" reward I get from that is: whenever I feel guilty for not having done much in a given day, I remind myself that by this action alone, I've saved my company several times what I would ever cost them.
Helps with self-esteem, but I don't think my company see it that way.
That's the difference between perm/consultant I guess.
Perm is more "we pay you so fix this", consultant is more the reverse "we need this fixed so we'll pay you".
I always find that dynamic hilarious because in a general sense permanent employees have more value than contractors/consultants as perms usually have a much longer tenure at a company (years and years, vs 6 months to 1 year).
However one thing I noticed after moving to the UK is that the culture is completely different here - everywhere I've worked here there are contractors who stay for years and years like a perm would. It's not necessarily a bad thing, but it does make me reconsider why anybody would ever be a permanent employee beyond a bit more job security (ie long term contractors definitely have to trust that they'll be renewed, even if they usually are it's always possible for the business to decide otherwise).
I once saved a company $20k in infra costs and saw nothing of it.
In fact the team was pretty upset that they'd budgeted that money for infra already and it'd have been better spent instead of waiting till next year to re-budget it.
I once talked AWS into a 5-digit refund for something that was our team's mistake.
It wasn't necessary though because we had a committed spend target to reach and we just had to figure out how to legitimately spend the money somewhere else. :(
> employees aren't liable for the company's losses and debts
Never been laid off during a recession or had your pay frozen and bonuses cancelled during a hard time? Employees risk a lot more than most stock holders by working for a company. On average, stock holders are way more diversified.
> Employees risk a lot more than most stock holders by working for a company.
Uhm no?
You apparently never been a business owner. Employees get their wages, and even can legally enforce them.
If a business go down, owners eat the losses and envy their employees.
I've been on both sides. Being a business owner is much riskier.
But wages are often also only a fraction of what they should be, especially for those working outside of the tech industry. But I suppose that's a separate issue.
Is the risk high enough to justify the ever-growing disparity between owners/C levels/investors/etc and the employees that get the work done?
What with all of the bail-outs through history, running an especially large company seems pretty much riskless. And hell, if you look at the history of technology, say games and game consoles (because I like retro games) the number of times a hugely successful product/project that netted 100s of millions of dollars was "not allowed" by the CEO etc but was hidden until it was too late (see Xbox etc) is super high. In addition to the number of decisions made by higher ups where the business swallowed a loss (particularly easy in larger businesses) is also high.
Imagine if Bill Gates and Steve Balmer hadn't been convinced/swayed to make the Xbox. How much profit has MS made from that? A fucking shitload, and have the guys that pushed it, or for that matter anybody in a similar situation (of which there are many) ever seen any of that success? No.
And we can't say "well the CEXs have the final say because they take on all the risk" they do, technically, but in reality when C levels screw up oftentimes it's just taken as a loss and things move on.
If you only take into account small businesses with single owners you are entirely correct. Mostly because large corporations are squeezing them to death with various forms of rent seeking. However, if your corporation is in a position to casually misplace half a million dollars, I don't think you are in that category.
LOL back in the late 90s and early aughts I had many friends in games. There were many times when payments didn't come in in time and employees were working for repayment promises.
> Employees risk a lot more than most stock holders by working for a company.
hmm, they can perhaps reduce the risk by not working for a company. They can just be stock holders or launch their own company, that way whatever may happen they will never get fired.
In order to become a stock holder or launch a company, you need to have capital. Capital is hard to acquire when wages are being actively suppressed by a cabal of employers. It is also hard to hold on to capital when your health care system is intentionally designed to strip away generational wealth from workers. Then to add a cherry on top, you lock the higher wage jobs behind an additional investment that can only be funded by non-dischargeable loans. Then you have the rent seekers, both literal and figurative. Landlords, insurance companies, toll roads, etc...
This isn't a risk the employee takes on as a result of doing business. It's a result of the company choosing to do this while also still making profits. So it's not really a risk, it's just mistreatment.
> Although there is a lot of capital investment in a job.
No there isn't. You haven't put up any capital when you join a job, and you aren't (generally) required to invest any as you go along. Your time is not capital.
Auto Loans, Student Loans, training, certification, internships, moving costs, and other small costs like special clothing are all investments made by employees to be paid for before or during hiring. In addition, many businesses require employees to buy tools and special equipment like boots and safety vests.
Correct, is way worse. Proletarians don't have any capital to lose. Capitalists do. So if they lose their capital, they can just be like the rest of us. If we (proletarians) lose our job, we risk poverty and death.
> On the other hand, employees aren't liable for the company's losses and debts, so it works out in the end.
How can anyone seriously type this? If you fuck up bigly enough, you will 100%—without fault—get sacked. Again, not even talking about long tails (bad economic conditions, layoffs, etc.).
This is under totally normal situations: if you lose the company money, you will be fired. As a bonus, you also lose unvested options or equity. These kinds of posts are exactly why engineers have garbage bonuses compared to finance even though they probably generate an order of magnitude more value.
> If you fuck up bigly enough, you will 100%—without fault—get sacked
which is to be expected - making a big mistake might not be something that can be forgiven and overlooked (depending on the magnitude of the mistake).
But you will not lose capital as an employee, since you did not put in capital to lose. Your time would still have been paid, up to the day you are fired.
Therefore, you obviously have no incentive to take on a risk that can result in a mistake (but which the reward you take no part in). You just do your assigned job, and whether it saves the company money or not, as long as you can cover your ass, you're golden.
Unless the company incentivize you to save money - for example, via a bonus through hitting a target or achieving some goal that was set.
> But you will not lose capital as an employee, since you did not put in capital to lose.
The conversation is a lot more complicated because there's an opportunity cost, you lose time (your time is finite, company time is infinite), you lose reputation, and so on. Besides, your argument is a bit weak as it's not like hedge fund managers put up the cash themselves, either.
My point is only that value-generators should be rewarded as such, and it's a bit weird that engineers are totally cool with not getting a piece of the pie.
> My point is only that value-generators should be rewarded as such, and it's a bit weird that engineers are totally cool with not getting a piece of the pie.
Software engineers are some of the best-paid labor in the world with great benefits and workplace conditions. They often receive equity as a compensation, even when the salary is still vastly above many other lines of work. They are absolutely getting a piece of the pie, and in much greater proportions than almost any other economic activity.
You may be discounting the value of capital, management, sales, and other roles in a successful software-related business.
The remuneration that labor and employees receive is never going to be in line with the value that they generate, precisely because the former group doesn't take any risk. They don't invest any personal capital and they aren't liable for anything. They can walk away any time, sometimes voluntarily, sometimes not. In return, they work fixed hours and get paid on a routine basis. The owners receive only what remains above and beyond all that, which could be great profits, just breaking even, or even losses.
> there's an opportunity cost, you lose time (your time is finite, company time is infinite)
Everyone everywhere loses time, because time passes whether or not you choose to do anything with it. Employees aren't unique among economic entities that they face opportunity costs.
> it's not like hedge fund managers put up the cash themselves, either
This is actually a good example to dive into. Hedge funds are typically paid "2/20", meaning 2% of assets under management every year whether or not there are any gains, and 20% of any gains above some benchmark. It's similar to, say, a commission-based sales role that gets paid a certain fixed salary and a percentage of sales they make. Whether or not 2/20 is "fair" is solely up to those who buy their services, since there is a competitive market of providers of fund management (the "employee") and providers of capital (the "employer").
And in some situations, the "employers" do in fact lose a lot of money, while the "employees" walk away; the limited partners of Melvin Capital, for example, lost many billions of dollars, all while Melvin Capital itself continued to charge the 2% management fee.
And within hedge funds itself, there are again employees who receive a stable salary and maybe some performance-related bonuses on top of that, versus the principals and owners who have personal capital invested. When LTCM blew up, for example, it's estimated that its owners lost $1.9B[0].
> The remuneration that labor and employees receive is never going to be in line with the value that they generate, precisely because the former group doesn't take any risk. They don't invest any personal capital and they aren't liable for anything. They can walk away any time, sometimes voluntarily, sometimes not. In return, they work fixed hours and get paid on a routine basis. The owners receive only what remains above and beyond all that, which could be great profits, just breaking even, or even losses.
The people at the top get an even better deal. They get given stock options, so they get the upside but not the downside. They can also walk away, but they'll get a big payout if they walk away involuntarily. They work fewer hours whether you're counting butt-in-seat time or making-efforts-about-work time (some people, bizarrely, compare the CEO's making-efforts-about-work time to the employees' butt-in-seat time and conclude that the CEO "works more").
> And in some situations, the "employers" do in fact lose a lot of money, while the "employees" walk away; the limited partners of Melvin Capital, for example, lost many billions of dollars, all while Melvin Capital itself continued to charge the 2% management fee.
You're flipping the categories. Being the "investor" can be a bad position, sure. Being the manager, the decision-maker, is where you can't lose. Concluding that that somehow makes employees better off than owners is ass-backwards.
It depends on how much you mess up. Mess up large enough as an employee and you can end up sued by your former employer. Losing a lawsuit is losing capital. A probably not comprehensive list of reasons an employer can sue an employee, not all of which are because of negligence or malfeasance: https://www.mylawteam.com/employment/can-an-employer-sue-an-...
You've completely missed the point. Businesses can lose money for all sorts of reasons. Owners have to eat the losses while keeping on paying salaries.
> If successful, Corizon's Two-Step would avoid a much wider range of liabilities than previous companies who've used it — not just injury lawsuits, like J&J, but the routine debts to vendors that companies rack up every day. If the company succeeds, it provides a "roadmap for eliminating virtually any unsecured liability owed by any corporate entity, regardless of whether that entity is solvent," Ian Cross, a Michigan civil-rights attorney who represents multiple prisoners who have sued Corizon, wrote in a procedural objection in April.
Their own company, using a handful of additional investors.
Or they'd sell to other investors, who are using a leveraged buyout, in order to get a better selling price than they would have otherwise. Meanwhile the new investors would take enough in income to cover the amount they put down, plus some, and then the company would eventually fail because it had too much debt.
In some way, I made Amazon get a lot of profit by choosing to purchase on their platform, and saw nothing from it, though I directly contributed to their profit and could deserve a %.
I saved/earned Amazon $25M/yr also back in 2012. Because the project that was supposed to compute price matching/most-favored-nation status for vendors was never actually implemented. Despite it being in the standard contract for vendors for years, no one ever noticed that we never adjusted prices based on it. My own initiative noticing the problem, my own design and implementation, as a L3(or whatever the fresh grad role level is), and I got zip for it. Big part of my reason for leaving the company. I didn't expect people to fall to their knees and worship me, but it seemed like a project that should be a big part of a promotion, but I was passed over multiple times.
