alexbeloi

If somebody robbed a bank and gave you the money, there would be an expectation that you would have to return that money, 'possession of stolen property' and all that. The fact that the money was given in the form of investment returns doesn't change that.

I think we can all agree that each investor should get their principal investment back. I don't think there's a good argument for letting people keep the impossibly high returns on that principal.

If you do the math, your "100x" credit comes out to 13.8% of your monthly payment in credit per hour of down-time.

That's not terrible or extravagant. The 100x multiplier is there only to make the credit reasonable, without it you would be getting chump change.

> the number of people that don’t mind (cold calls) statistically out number the ones who do. We have tons of data to support this. I highly recommend evaluating this as a strategy in any B2B customer acquisition program.

I don't doubt it's a good strategy, but I doubt most people actually "don't mind". What kind of data do you have to support your claim?

If you replace "don't mind" with "tolerate", I would find it more believable.

>I can honestly say that stimulants make a vast, objective difference to my work output and quality

What objective measures are you using to evaluate yourself? are you in the top 1% of earners for your experience level, position, city?

Anyone considering using this for continuous hyper-parameter optimization in machine learning should be warned that the automatic Lipschitz constant optimization (the novelty of the method) will very likely not work well with a noisy objective function (any stochastic learning algorithm).

edit: Any grad students looking for a paper idea might consider extending this result to allow stochastic objective functions satisfying a stochastic Lipschitz condition.

You're not wrong, the objective is assumed to be Lipshitz which is continuous.

Maybe, my point is that isn't not a 'win' for the laborer like you seem to suggest, that maybe 50% of the people losing their jobs get to work better jobs. The people losing their jobs that are lucky enough to not be laid off are more likely be relegated to lower paying and lower skill inspection jobs rather than the higher skilled machine maintenance/development.

IMO, the most likely outcome of your scenario is: 19/100 of the workers become visual inspection team, (maybe) 1/100 of the workers becomes a manager for visual inspection team, 5 external machine specialists get hired, 80/100 of the workers get laid off and 75/80 of those will continue same line of work for even less pay (supply/demand) and 5/80 will burn through any personal savings/credit to retrain themselves with the hopes of becoming one of the '5 external machine specialists' hired somewhere else.

These numbers are definitely made up by me based on my own world view. From the original 100 laborers, 94/100 of the original people are worse off than before and in exchange we have 5 (maybe 6) new skilled positions.

I'm not against automation or productivity or higher skilled work, I'm personally invested in all three, I just think that if we aren't aware of all the consequences that we are likely to be blind-sided by a resurgence of Luddite riots.

>In the case of the Google keyboard I use, when I finish swiping it doesn't "present a list of possible words for selection" - it just chooses the "best" one.

IANAL. They probably dropped the ball in writing "presents it to user". If they had stopped at "generates a list of possible words associated with the entered part." There would very likely be a sub-process that generates a list of possible words in all other implementation, that sub-process should be violating the patent regardless of what the user sees.

>Going from 100 workers at a factory to 50 workers means the 50 workers remaining have harder and more valuable jobs.

This is not necessarily true. It could mean turning 100 people doing skilled labor and turning them into 50 workers doing visual inspection (QA) at likely lower pay.

>Workers aren't paid what they deserve, they're paid as much (or as little) as they're able to negotiate.

Deserve is a tricky notion. The capitalist will argue that workers deserve precisely their market rate, e.g. what they are able to negotiate (collectively or otherwise).

Stability is a more appropriate term.

Japanese culture values stability very highly. People in this thread believe this kind of stability results in low economic mobility. In the US, people associate high economic mobility with meritocracy.

There's a couple leaps to be made in both associations, but I think that's what the gist is.

You could also spend months of your life learning the intricacies of measure theory, but then every measure in probability is a positive sigma-additive unit measure.

One can learn the necessary topology, analysis (etc.) in the relevant places (and the relevant depths) that they come up.

I would amend as follows:

Skip abstract algebra, topology and analysis. If you find yourself in the same room as a number theory book, walk away slowly without making eye contact lest it cast a spell on you.

It's a shame how data hungry DRL is (even when compared to DL), but the DRL framework encompasses all the standard classification/regression tasks and also includes decision making, planning and pretty much anything else you can think of.

Model generality and data efficiency are in an inverse relationship and a lot of research has been in moving up or down this hyperbola. On one extreme, tailoring models to specific use cases/datasets/environments, on the other end transferring learning across domains. DRL is stuck pretty high on that generality end. Some breakthroughs seemed to have moved progress to a higher level curve, optimizers (Adam, DQN, TRPO) have gotten better which helps everything in general, core structures like CNNs or memory cells, which seem to be somewhat universal (or our best guess yet), but there's still something fundamental that seems to be missing. Or maybe this is all there is and we just need a computer with a richer/higher-resolution sense field and the flops to process them.

That is the cost of doing business in China, and it (doing business in China) is voluntary.

If you are an employee, your cost to your employer is likely to be close to 2x your salary.

I'm no expert but this seems like a good case for doing random subsampling during the test.

IMO fast probabilistic total coverage test is better than no test or partial coverage.

The most surprising thing to me is that they can parameterize a nontrivial section of the implementation space of a function, or that such a section exists that hasn't been optimized away by the compiler.

>https://networkengineering.stackexchange.com/questions/45602...

Cable internet is a broadcast network, every signal from your ISP's exit node is sent to every client attached to that node in your geographically local region (I think this can be a tens or hundreds clients) and vice versa. Your modem only looks for frames that are designated for your modem's MAC address, the rest are ignored but still received.

Same as channels in broadcast tv, everyone is always receiving the signal for all the channels, but they only 'tune in' to their channel. The difference is that your internet modem is only allowed to tune into one channel (designated by its MAC address). I believe each modem has a unique key for modulating/demodulating its signal, and this is used to establish secure communication with the ISP.

In the book "Computer Networking A Top-Down Approach" the following is written: "One important characteristic of cable Internet access is that it is a shared broadcast medium. In particular, every packet sent by the head end travels down- stream on every link to every home and every packet sent by a home travels on the upstream channel to the head end. For this reason, if several users are simultaneously downloading a video file on the downstream channel, the actual rate at which each user receives its video file will be significantly lower than the aggregate cable downstream rate. On the other hand, if there are only a few active users and they are all Web surfing, then each of the users may actually receive Web pages at the full cable downstream rate, because the users will rarely request a Web page at exactly the same time."

tl;dr you already probably broadcast your modulated bank info (if you're on cable).

>additionally, for virtually all stores on instacart.com now, we are simply passing on the pricing that the retailer has given us (which may or not be marked up from their internal prices -- but we do not control)

IC is responsible for signing the contracts they have with retailers, IC is also the most visible partner for the end-user. As you can see from this thread, attempts to kick the blame down(up?) the supply chain will only fall on deaf ears.

>This doesn't work, because candidates can (and apparently do) lie.

I have not found this to be the case. I ask people about specific projects they have posted and sometimes look through the code together and ask them about the time-space complexity of function they've written.

Maybe you're just encountering test-anxiety.

>I don't want to be condescending, so if you have a better suggestion, I'd love to hear it.

Almost every interviewee has a git(hub/lab) account with code they're written, that should suffice to determine if they can FizzBuzz. You could also ask them for a coding sample prior to the interview if they have no public code.

The solution is simple, nobody does it because it asks the interviewer to do more than skim the resume before walking into the room. I think this comes from the perception that the interviewee's time is cheap compared to that of the interviewer's.

In a perfect world, neither group would care about either of those things.

The corporation is run by people and people favor short term results when they expect to be in their position for a short period.

Just as it's far less common for the regular working (wo)man to stay at the same company for decades, it's less common for CEOs to do so as well.[0]

If a CEO/president expects (through loyalty/control of the board, or otherwise) to be at the company for a long time, they will naturally make long(er) term decisions.

[0] https://en.wikipedia.org/wiki/List_of_Hewlett-Packard_execut...

For (2), developers aren't the ones making the decision to deprioritize bot detection. The decision makers don't care because the ROI on detecting/removing bots is likely negative, whereas using those ML people to improve ad targeting and engagement is much higher.

Because investing in coins is non-productive, those coins don't generate anything besides demand. If one believes in the technology, one should be investing in companies working on expanding blockchain tech. Also blockchain != coins.

A military base buys labor, technology, services, raw resources and produces defense/security. You're conflating what the military base spends with what it produces. (the point of the argument against GDP)

Just because I spend money on something, does not mean that I produce that thing, even if my spending increases demand for said good. That results in double counting production, which was already counted once by the seller.

In tech, when your employer hires you as an engineer, they hope they're getting 10x your cost back in value each year, at a minimum a startup needs to get 2x return to cover debt obligations with growth, on average they get 3-5x. You can get a rough estimate (with profit or earnings as a proxy for capital growth) by looking up a company's (or sector's) profit per employee.

These are averages and not reflective of engineering roles.

Walmart: ~5-6k profit per employee (Q3) Exxon: 176k profit per employee (Q3) Facebook[0]: 1.7M profit per employee (Q3) Apple: 522k profit per employee (Q3)

Internet Services & Social Media sector average: 497k profit per employee Internet Services & Social Media sector average: 2.1M revenue per employee

A competitor can/should/will try to steal you from your current employer with a higher compensation if they can extract good surplus value from you, relative to anyone else. The last part (relative to anyone else) is the crux of the situation, everyone selling their labor is doing so at a massive discount because of the competition in the labor pool. You can't actually demand anything close to your value because somebody else will undercut you.

The people/firms that can demand the highest percentage of their value as compensation are those with the most unique skills, those for whom few alternatives in the pool exist.

[0] https://csimarket.com/stocks/FB-Efficiency-Comparisons.html

Your value is what you earn for your company. The difference between your value and your compensation is the surplus value the company collects from your labor.

The company's goal is to maximize aggregate surplus value. Salary and wage based negotiation is a zero sum game until you add commissions, stock options, profit sharing.

post-WW2 era