SomeStupidPoint

> No ninjas, rockstars or brogrammers, please; just nice, caring humans.

I like how in a post about inclusive job ads, they still wanted to show off that they Exclude The Right People.

Their cultural based discrimination is obviously correct and should be applied universally; previous cultural based discrimination is a crime we should all strive to avoid!

I actually enjoyed the post (and agree with most of the points about writing job postings) and don't mean to be overly critical, I just think it's an interesting chance to reflect on how easily these habits come to us as people. (And so some of the language above is dramatic for effect.)

I just wished they called it something besides "debate".

It made it cannoying for those of us who were actually trying to organize a club around debating issues.

I'm not sure I understand your question, but the one that most pertains to ECC would be preventing corruption of the routine performing the checks --

It varies greatly by the method you use to ECC in software, since they all feature different mechanisms.

In the case of the one I had in mind, the kernel is relatively small (so errors are unlikely), and once loaded, it's stable against bit errors. There are probably better answers, I'm not an expert on software ECC.

(Tampering is a different issue -- if they can tamper with the program, ECC won't save you from them writing garbage to memory. So that a hack which can write memory can bypass memory safety isn't really interesting.)

> Or even better how type systems could work around data integrity errors in RAM without ECC.

You realize that you can do ECC in software too, right? You just need the statistical bound on a double fault to be similar to that of the hardware case.

Type systems can be used to do this by defining an interface for the data, letting you swap between normal datatypes and ECC ones seamlessly. (And in practice, you only need to define the base operators for your new type.) You can do this with ducktyping ECC versions of objects in say... Python. But the type system helps ensure that you've built the full set of components necessary and it properly links. You have to do this with ad hoc IDE tools in Python.

Similarly....

> The defect was as follows: a one-byte counter in a testing routine frequently overflowed; if an operator provided manual input to the machine at the precise moment that this counter overflowed, the interlock would fail.

That actually sounds like exactly the situation that a type system would be used to prevent -- partly because from a type theoretic perspective, that whole situation sounds like an obvious defect.

Further, the result was because of an obscure combination of key strokes that went untested. The use of total functions (enforced by the type system) lets you verify (by computer) that no combination of input can transition to a bad state.

So... Did you intentionally pick two cases where type systems are the obvious solution to the problem?

> No, but not liking type systems might be an intrisic property of some programmers

I wish they'd stop writing shitty, ad hoc type systems into our production code, then.

There's actually a few Buddhist sects that would be aptly summarized as "Treat people better so there are fewer reincarnations."

Just a quick comment for those taking this as advice:

You probably wanting something in the 40:60 to 20:80 THC:CBD range if you're looking for less intoxicating pain management. In genearl, CBD > THC will lower the intoxication (though not completely remove it!), with the higher ratio lowering it more. However, having some THC will make it much more effective at pain management. (Too much lone CBD veers into being psychoactive-esque, as an anti-psych.)

(The interaction between cannabinoids and you is complex.)

You are correct in theory.

However, the present system is so bad that it routinely fails at (eg) streaming Netflix even though this uses less than your advertised bandwidth (by a wide margin).

So at present, it would just be a matter of automatic documentation of failures -- and the situation would be much improved if we were up against the limit you propose.

> Their financials look fine, they have over a year's budget in reserves.

I think the question is if they actually need a $65mil operating budget -- or one that's growing 15%+ per year.

Especially since they're giving away a sixth of their budget in awards and grants! At that point, they're taking on a secondary role as a financial institution funding projects in addition to their primary one. If they cut this, they'd need only $55 million and would have almost two years of reserve.

(We can also talk about if salary creeping up from a third to half the budget over the last ~5 years is appropriate. But that's a smaller issue.)

Isn't that what kills most charities that don't starve?

Could you outline some of the problems you think current research isn't addressing?

Similarly, regulated drug dens (like regulated brothels) are a perfectly tennable idea. There are middle ground solutions between "ban everything!" and "sell it everywhere!"

It's surprisingly hard to OD with a clean supply, measured dosing, and emergency services on site. I don't think that clean, supervised drug use is actually any more dangerous than say... skydiving or bungee-jumping.

(It's actually hard to fatally OD with just one vaguely sober person an nasal naloxone, but that's more specific to opiates.)

Considering that they're making a specific claim about their usage and then using the access you provide to gain additional data in excess of what you approve, does this constitute hacking (and hence cybercrime)?

At first blush, this seems indistinguishable from social engineering installation of an application that steals data -- massively done against thousands to millions of victims. I suppose that would make this website one of the larger hacking groups in the world.

Microsoft has recently switched to an unhelpful, enraging chat bot for their help service and hidden their actual contact information.

The whole point of electronic computers being able to perform algorithms was to replace the work of computers (the human kind) in huts filling out sheets according to strict rules.

They're literally an attempt to replace humans as the mechanism by which rote paperwork can be done -- that they're used this way by business isn't terribly surprising.

So algorithms aren't the "new bureaucracy", they're the new way of doing the old bureaucracy -- as they were since inception. This is a feature, not a bug.

Wait, is your contention that the US can't have a series of worsening problems with drugs over the last 40-50 years as wages have stalled and the economy rotted?

That seems... Unreasonable.

If a problem is growing, every iteration should be the worst in history.

From [1]:

> Rather than putting it immediately into the market, he waited and invested after month-end January 1993—that year's monthly low point for the S&P 500. At the beginning of 1994, Peter received another $2,000. He waited and invested the money after March 1994, the monthly low point for the market for that year. He continued to time his investments perfectly every year through 2012.

That's hardly "perfectly timing the market" in the usual sense. I'd consider "perfectly timing" to be sell at every peak and buy at every trough, so you're always holding stocks when they're gaining versus the dollar and holding dollars when stocks are losing versus the dollar.

It's impractical, but it's a much better result -- the theoretical limit to how much you could have made with perfect knowledge of all the pricing ahead of time.

That analysis, by contrast, only looks at what happens if you have a little perfect information -- and the answer is "not much".

So if that's your issue, why not try to solve it?

Try having less of a routine, not in terms of schedule but of habits --

Entropy is the crux of creativity.

I've considered making a "life randomization" app to add perturbations to my own habits. So if you solved your own problem, which is also one of mine, you'd have at least one user.

Just a thought.

> Constantly being assailed like that takes its toll, and can lead to those people withdrawing from social media, removing their point of view from the discourse, and emboldening that of the bigots.

From my understanding of social media, it hasn't been workable trying to stop outright racist comments being hurled at those marginalized people -- so I just can't agree that not actually addressing the cause so we don't hypothetically do less damage than is actually being done right now by the problem is a workable idea. That's the kind of squeamish bargaining done by people without the stomach to tackle complex or hard topics: "Can't we all just get along without it being messy!"

Your complaint is like saying that surgery hurts, can't we just let the infection fester? The surgery doesn't even hurt as much as the infection! Why would we do that?

This is, for the moment, setting aside that we could work with technology to do better at time/place of those questions and filtering so the stream isn't overwhelming for (or even involuntarily sent to) any particular marginalized person. And so at a technical level, you're solving the problem in an inefficient way (trying to legislate human nature versus changing how we route text online).

We could (and I have) had whole discussions unpacking the thinking around that quote. There are multiple levels of racism (and other -isms), misunderstanding, and legitimate concerns that could be being expressed. Assuming that it's just "a false premise" (though, there's many false or questionable factual premises -- eg, program targets mothers not single parents; most single mothers have children with black fathers; support programs have anything to do with child support) without really investiating what else it might mean is too dismissive -- it's shouting "tree!" then being proud without ever seeing that there might be a forest.

I takes time, effort, and empathy to try and unpack the real meaning behind questions like that -- effort that we don't invest when we allow ourselves the shortcut of just calling them "racist" to end the conversation.

I don't think we can move forward as a society -- and are even regressing -- because we've stopped unpacking what the other person is trying to say and started short-circuiting that process of human interaction with thought terminating cliches. It sincerely scares me how many people derive great joy in that kind of unthinking censorship.

Has that method successfully prevented bigots from collaborating, growing, and exerting political influences -- say in elections over the past year?

I'm not saying we give them a megaphone, but if bigotry is prevalent (and it seems to be), then the best antidote is to expose it and address that directly.

Why?

Simple. Bigots are bigots for reasons -- even if they're not (in our esteem) good ones. You can't address the symptom (bigotry) without addressing the cause (those reasons).

Your method only leads to wounds festering under bandages, and not the exposure and cleansing we need for healing.

Loaded ones, eg:

"Why should I pay for taxes to support single mothers just so black men don't have to pay child support?"

Now, I think we should still encourage them to ask those kinds of questions, because I believe in an open and liberal society and that the best reaction to poor thinking is to address it directly, that the answer to speech I don't like is more speech. But it's pretty easy for some really racist opinions about others to hide behind questions like that.

Of course, there are also some uncomfortable truths in those questions which cause us to silence them, such as -- "Are the disparaties we see in racial groups purely the effect of lingering racial bias, or is there a modern cultural component that happens to align with racial groups and perpetuates the inequality?" (And in this case -- the more general, is there a disparity in the fatherhood rates that correlates with racial grouping?)

Ed: I think the people downvoting proved someone upthread's point -- I literally provided an example I've heard of such a question and got downvoted for providing facts to a discussion.

Also, there are lots of specific configurations of cities for which an optimial solution can be found in a tractable amount of time -- it's only the general case that is the problem, and there's some debate about how "degenerate" real world examples are. In "well-behaved" cases, a good approximate answer will be the answer; in almost all cases, the differences between a good estmate and the correct solution is small relative to the total path (and so the difference is of limited importance).

I often find that optimization problems suffer from trying to optimize the general problem rather than using the constraints present in the real world.

TSP is similar to Turing machines and the halting problem -- in practice, it really doesn't matter.

They actually have some features that aren't available in other clouds (well, AWS/GCP/AZ) -- if we weren't so locked into AWS, we might've tried them for that reason. We wanted to know if we could run our own hypervisor in their cloud and access VT-x instructions. (And there were other companies at their talk that were definitely hitting up sales for trials; the "baremetal" aspect intrigued a lot of people with compliance reqs or complex cloud appliances. Dedicated bandwidth was another.)

I think later clouds have advantages and not just missing/worse features -- they got to design their architecture after seeing others try, and definitely learned some lessons.

I don't even get why they would. It's not like the time I have to type code is my limiting factor on productivity.

Past a certain point, I'm not solving issues -- I'm creating them. On any horizon over a week, I'm better off sleeping, eating, exercising, etc than trying to smash every hour into working. Keeping my brain on a routine and properly nurtured is more productive than crazy hours, by a wide margin.

But I guess looking busy can be as important as results.

Contracting instead of developing in-house capabilities is completely destroying the DoD and American military effectiveness. The corruption around that is posing a real, impactful threat to national security.

Fuck these people and their "free market" lies as a cover for outright theft of public funds.

It's not just in the cyber domain that this is a problem, but the cyber domain is one in which the corner-cutting, half-assed nature of the corruption is most visible because the damage is most easily exploited by foreign powers.

They actually sell your psychological well-being to advertisers after binning you based on personal data.

That's Facebook's business model (and what made them a useful stooge during the election).

Slightly modified version, which uses some of the bounds adjustments suggested by other comments:

    int n = 100
    vector<int> p(n);

    void sieve()
    {
        for(int i = 2; i * i < n; i++) {
            if (p[i] == 0) { // Could probably just !p[i]
                for(int j = i * i; j < n; j += i) {
                    p[j]++;
                }
            }
        }
    }

    bool is_prime(int k)
    {
        return p[k] == 0;
    }

> in other words he's not doing the test that something is prime, and crossing out the multiples of composites as well.

This doesn't change the outcome, since the multiples of composites are also multiples of primes and never primes.

> There are other problems as well, he adds one each time to p[j], but tests to see if it's ==1, so that won't work.

You're wrong here -- testing equality to 1 for the number of times an entry in the array was hit is exactly testing of primality post-sieve. If it was only touched once, then it has nothing that divides it (since the only cycle it was incremented on is the one starting there).

> it's certainly not the actual Sieve of Eratosthenes for any one of several reasons

If you initialized the array entries to 1 and started indexing your inner loop at j= i * i, it's exactly the sieve (modulo multiple strikeouts of removed numbers).

> It's tricky to explain

This reasoning is fairly simple:

Suppose we have a composite number K that is less than P^2.

Then K = A * B, for some A and B not equal to 1, (which is the definition of composite). Let us say that A is less than or equal to B. (If this is not the case, switch the labels.)

Then A must be less than P:

If A is greater than P, K = A * B > A * P > P^2. But we chose K such that K < P^2.

If A is less than P, then K will have been marked during the cycle of Q, a prime divisor of A, which will occur before P. Therefore, we do not need to examine any numbers below P^2 -- they are either prime or have been marked.