Obviously the SSC readership is a very long way from being an unbiased sample of (say) the whole world's population. No one would expect it to be, and in fact that's the point here: a survey of people who are obviously unusual in some respects (whatever combination of quirks turns someone into a likely SSC reader) turns out to be unusual in another respect with no obvious connection, namely having substantially more firstborn children than you'd expect.
Whatever it is that makes someone more likely to read SSC, it seems like it's probably a combination of things that surely can't correlate with birth order (e.g., being a native English speaker) and, broadly speaking, personality traits (e.g., being interested in the sort of thing Scott writes).
So the results show evidence of a link between birth order and personality, and (from the survey results) apparently a strong one. Which is interesting if true. And all of this only works because the SSC readership is far from typical of the population of a whole.
So, again, what do you mean by "skewed"? And why is it a problem?
His writing indicates he believes his readers are "more intelligent", and he believes that first borns are "more intelligent". So yeah, if you were a second born, who was more intelligent, you'd dismiss his blog as BS, and would not be counted among the survey respondents.
For that to work, you (our hypothetical more-intelligent second-born) would need to:
1. Have seen Scott express the opinion that first-born children are more intelligent.
2. Interpreted that claim in a way incompatible with your own experience of being an intelligent second-born (e.g., taken Scott to be saying first-borns are always more intelligent).
3. Been sufficiently offended (or otherwise unimpressed) by this to stop reading his blog when you would otherwise have been happy to read it.
For this to explain the apparent firstborn bias in Scott's survey respondents, thousands of potential readers would need to have done this. So, how plausible is it?
1. I've been reading Scott's blog for years and don't remember ever seeing him say anything like "first-born children are more intelligent" before this post we're discussing now. That doesn't mean he never did, and I'd be extremely unsurprised to find that he did -- but it does suggest that if he did it was easy to miss.
2. The distinction between "on average firstborns have an IQ one point or so higher than non-firstborns" and "firstborns are always smarter than not-firstborns" is not exactly subtle.
3. Well, anyone can get upset about anything, but this really doesn't seem to me like the sort of thing that would make thousands of people swear off an otherwise interesting blog in disgust.
I'm going to rate it very implausible. No way is anything remotely like this a non-negligible fraction of the explanation for the apparent firstborn bias in Scott's survey respondents.
This sticker only works against this classifier. If you start changing the algorithm, you'd need to change the attack to match.
If you think you can write a better image classifier by first segmenting the image before using ML, then I encourage you to get your own computer vision paper published and see how that works for you.
That's foolish. The lead up time to become an expert at deep learning: Probably 6 years minimum.
No one hires people to do deep learning with 0 years experience.
Very few people with the skills you assume by default (basically anyone with programming skills) would be convinced to leave their cooshy jobs, but plenty of people with no skills or prospects would.
I have literally no idea how you could have reached the idea that you would be an indentured servant/bill collector. Is this some sort of post-truth propaganda?
I understand why people are greying out that comment. On the other hand, the idea that municipal services lead residents to literally collect payment from their neighbors is such a novel complaint that I'd love to hear more details.
eldavido- can you give some more information about any news stories that cover this happening?
I'm not saying anyone is collecting bills from their neighbors. I think people are reading my comment a lot more literally than I intended.
What I'm saying is that, we have large-scale utilities like power and water companies that are able to provide individual, metered service to households over a geographically distributed area. In order to to this, there is a large amount of unpleasant, schleppy work required that most analyses of smaller-scale alternatives conveniently sweep under the rug.
Either we assume (a) another large-scale operator will provide broadband, which will probably be something like Comcast (maybe not), or (b) it will be done at a smaller scale, say, a company for 1000 residences. Does (b) actually exist?
Chattanooga's municipal broadband exists and delivers speeds as fast as any in the U.S. Salisbury, NC has a municipal broadband service that was grandfathered in before the state law banning municipal broadband.
As far as b: I live in a small town that started a municipal wireless service after Wheeler's FCC tried to override the state law banning it. Fast, affordable, symmetric. I didn't attend the meeting but am certain a clear and detailed cost-benefit analysis was presented and voted on (as well as being made available to the public per state law).
Interesting. Care to tell me a little more? How many people live in your town? Is it a socially cohesive place? What would people qualified enough to run the ISP do if not work at the ISP?
I always think of that last point (what else would people do) when people mention how "the Soviet Union had such great math teachers OMG WHAT IS WRONG WITH THE US". The US has SILICON VALLEY, hedge funds, world-class universities, etc. - US primary/secondary ed has to compete with these places for talent and frankly, they often lose.
I have a hard time believing that a small operator could offer anything that stacks up against Comcast product-, service- and price-wise, but I'm willing to be shown I'm wrong about this. Maybe I shouldn't be so cynical ;)
All I'm saying is that, whenever people complain about Comcast, it's inevitably about some problem with either their network, or customer service.
Comcast is a huge, scaled organization that deals with "the public". They deal with people who don't pay. They follow up when bills get lost. They deal with people who forget to return their routers. They provide level-1 tech support for people who accidentally switch their modems off and call to complain. They handle shitty premises wiring problems induced by home DIYers who crimp RG-6 connectors with pliers. They handle support cases where people spill Coke on their routers and complain that it "stopped working".
All I'm saying is, it's just incredibly naive to look at the problem of an ISP as just "network operations". Like, we can build/operate a better network=profit. That's an incredibly naive view that neglects all the ancillary work of premises provisioning, bill/payment management, 24/7 customer support, pretty much the entire "customer" side of running a large-scale retail ISP.
I mean really, what is "community broadband"? Are we just going to magically assume that, because we aren't dealing with Comcast, all of this work will somehow magic itself away? I don't get it.
I think "falls apart" is overreaching. If the claim was that 21 states were targeted by russian hackers, and only two states repudiate that, that's less than a 10% error rate.
Keep in mind that state governments want to make sure they aren't deemed incompetent. But anyone who knows anything about "the cyber" knows that there are tons of legacy systems at both the state and federal level which are wide open to possible attacks.
The suggestion that an adversary could "scan" the networks is totally within the realm of plausability. This is the first step of a cyber attack, and while an individual state may be able to conclude that they weren't successfully attacked, it doesn't mean they weren't targeted to begin with.
Pretty skewed results.