sjbase

I'm not sure "the startup is disappearing" is the only explanation for this data. Isn't it possible that companies are just lasting longer than 2 years?

What if there was a larger percentage of age 0-2 companies '85 because back then, so few were making it past 2? Maybe easier access to capital has increased survival rate (or at least slowed down dissolution by a few years).

This explanation also fits better with the massive influx of cash into VC over the last ~10 years.

To me, the hidden crux of this whole issue is "can this system be as effective to deal with as a human who is good at their job?"

If that condition is met or exceeded, none of this voice stuff will matter. People won't care if the system discloses it's nature, or sounds creepy, if their airline reservation change (or whatever) goes abnormally well.

If that condition is not met, which IMO it won't be in the near future, then we're just adding attempted manipulation and potential for abuse on top of an already bad UX.

Question, do you use any of the Elixir linter/debugger/highlighter plugins? Have any recommendations?

I'm early on in evaluating ElixirLS and vscode-elixir but have run into some issues with both.

I find it suspicious that every one of these causes is a form of fiscal or monetary policy.

Keywords that don't exist in this article: Job, Workforce, Education, Technology. None of those even get a place in the top 10?

OneChronos cofounder here. The NMS order protection rule certainly doesn't make things easier, but periodic auctions have a unique compliance and fairness advantage there.

The question most continuous limit order books face is "What is NBBO right now?" That question can't be answered thanks to the speed of light being finite. But because of the way our auctions allow time for orders and market data to arrive at the matching engine, we can more thoughtfully construct a consistent view of "What was NBBO 1 millisecond ago, at the moment we struck our auction?" This is a much easier question to answer, and also has the effect of smoothing out network jitter, data gaps, etc.

Also, the auctions are combinatorial but are computed across linear components (stock symbols). NBBO applies to symbols individually, so it can be taken into account as a standard constraint on each symbol.

I don't think they really have a choice here. If you think the SEC rules on "is X a security" are broad, take a look at FINRA's rules on advertising and recommending securities.

The crypto litigation mass is approaching its Schwarzschild radius... By acting as a conduit for these ICO ads, Google etc. are making themselves part of the accretion disk arond the black hole, teeing up a decade of subpoenas, complaints, etc.

> It took the truth about six times as long as falsehood to reach 1500 people and 20 times as long as falsehood to reach a cascade depth of 10. As the truth never diffused beyond a depth of 10, we saw that falsehood reached a depth of 19 nearly 10 times faster than the truth reached a depth of 10

The study attributes most of that force multiplier to novelty and surprise, where falsehood will always have an advantage; important things that actually happen are often not surprising. Maybe people who wrap truth in clickbaity headlines are actually heroes, trying to even out the playing field... I never really thought of it that way before.

Nice work and grats on getting to the elusive "done" stage :)

Income tax is a big piece missing from the expense model. I'm guessing it's easy to add that line item, but it even if so it would be cool to have it computed automatically since its mostly just a function of {domicile, entity_type, pretax_rev}.

I think you'd find the philosopher Jean Baudrillard really interesting [0]. He wrote extensively about how humans create structure/models/frameworks around everything that approximate the "real" world into something we can work with. He calls them simulations, and i think he goes as far as to argue that it is in fact the "real world" underlying those simulations that doesn't exist; or ceases to exist once it's been simulated.

0: Excerpt from his book Simulacra and Simulation - https://web.stanford.edu/class/history34q/readings/Baudrilla...

Exactly true for stocks, but I don't think Apex or any equities clearinghouse is clearing cryptocurrencies (exception: clearport for CME futures). So we're left guessing as to whose balance sheet holds the crypto assets.

> My guess is the people who have such black and white views haven't ever been tasked with marketing a product in a true professional capacity

This is so, so true and in such a broad way. I've had this thought gnawing at me for ages that these people who are doing seemingly evil things (think James Comey, Ajit Pai) have likely been tested in ways I never have. Who knows how I would handle being in their situation? I don't think it excuses or absolves true wrongdoing, but it does give me some sympathy or at least reasonable doubt of malice. This an elementary concept, but it never seems to get mentioned.

To be clear, I'm not one of the people who categorically hates your profession. I'm an advocate of privacy, I avoid most types of social media, use uBlock, etc. But I can see it being hard to figure out where the line is, and how to not cross it. I've never worked in the consumer/media/ads world and had to face the "what to track" dilemma head-on.

Intuitively I'd agree that "best intentions of self vs. others" is probably the mechanism behind this type of cognitive dissonance. What I find harrowing is that in all likelihood, this is true for myself as well in ways that I'm not aware of!

Your point about denial also reminded me of an article a read a couple years ago, attempting to explain why humans deceive themselves. The soft conclusion: we deceive ourselves in order to better deceive others. In this case, we convince ourselves that we have the best intentions in order to convince others of the same. I can't find the exact article, but this is similar and refers to the same research by Robert Trivers. It's an interesting read: https://www.scientificamerican.com/article/living-a-lie-we-d...

This thread is a really interesting example of how easily humans can simultaneously hold conflicting beliefs/opinions. I'm gathering that a lot of developers and businesspeople here:

a) are very concerned about collection of their own data

b) derive material value from Crashlytics, Mixpanel and other "tracking tools" for their work

It's tricky to reconcile those two ideas.

Quote in the article, directly from Crashlytics:

> "and get insight into your users, what they’re doing, and inject live social content to delight them."

It's also analytics, and ads.

Potentially. But then for biological, you'd also need to include whatever system generated the ATP, oxygen, electric impulse, etc to power the muscle, right?

IMO, the crucial piece of information: how much pneumatic energy is needed per unit force for one of these, compared to that of more naive designs.

This appears to be an amendment adding $100m to a July 2017 filing, which was for ~$350m:

https://www.sec.gov/Archives/edgar/data/1181412/000118141217...

Would you mind elaborating on which aspects of the language reflect poor design? And what was done well for that matter?

I have no strong opinions on Solidity, btw - just generally interested in people's views on developer experience.

It's really unfortunate. There's a great book called Metaphors We Live By [1] that digs into this same observation, using "arguments" as an example: you "attack" another person's argument, you "defend your point." You either win or lose. We could have just as easily chosen arguments to be a dance instead of a war.

A lot of narrative here in America eventually turns into good vs. evil, and War on X is an example of that. I've heard people attribute this to anything from religion, to the revolutionary origin of the country, to military-industrial conspiracy.

[1]: http://press.uchicago.edu/ucp/books/book/chicago/M/bo3637992...

Indeed, however he's also suggesting that said evidence is missing under very suspicious circumstances:

> "That the most important witness in the case [Levandowski], somehow his own files got overlooked. That's very suspicious."

Is he not implicating Uber in that, or at the very least Uber's legal counsel?

True. Small comfort, but if you're a purchaser of credit rating data, you may have some flexibility in who you purchase from.

Nope, I don't wait for courts when I think I have enough info to form my own opinion. I'm just as outraged about the Equifax breach as anyone else, and would never do business with them in any way. The key is, there's more evidence every day of negligence: the obvious authentication issues on their site, the "SSN API", the 4+ months that they didn't install the Apache security update, etc.

For me personally, the trading issue is a related but separate incident; one hasn't crossed the same threshold of clear evidence.

Definitely - the "very suspicious circumstances" part makes sure of that.

Just a friendly reminder that due process is a thing in the US. Let's allow some time for the texts, emails, testimony, etc. to come out.

Everyone seems to have reached conclusions about this on very suspicious circumstances, but almost zero facts.

Interesting that under the Chalmers theory, there's plenty of room for computers as we know them today to be conscious. That "critical point" where quanta become qualia, if it's purely a matter of complex arrangements of bits, could have been reached by now. I'm sure I'm not the first to put this forth, but maybe your laptop undergoes some form of experience like the "phenomenal mind." One where it perceives beauty in complex patterns emerging in sequences of instructions. Or pleasure from a hard computational task executed with low time complexity.

I suppose that we haven't observed any of the effects of consciousness, i.e. expression of choice, could be an argument that there is no such consciousness in a computer. Or, it could just be that computers are conscious but lack any tools to exert free will. Their consciousness is "read only," so to speak.

Well, the very first assertion "Banks make currency" is not accurate in the US (the context of the article). Currency is "made" by a combination of the US Treasury and policy set by the federal reserve. The fed is a bank of sorts, but in no sense is a bank like JPMC creating currency.

> "This is like someone who makes a salary of $60,000 having suffering loss of $6,000 to $12,000."

This line had the opposite of the intended effect, for me. Someone with $60k salary, a $12k loss could take years to recover from if they have expenses like college debt, rent/mortgage, and kids.

I dunno, some might say that's poor financial planning, but it's reality for a lot of people. And de-referencing the analogy: most US cities are not exactly financially rock solid either.

The key statement being "sensors [...] that they think will enable level 5 autonomy." One or both of GM and Tesla could easily be wrong about that.

But you're completely right... it's like saying my Macbook is the world's first strong/general AI, once the software's ready.

Wow, yeah you're right. I gave the feasibility analysis a more thorough read and it seems like they're exploiting the amplifier inside the mic itself to down-convert the frequency on portions of the signal. I don't fully understand what nonlinearity is being taken advantage of, but to as an acoustics layman this seems pretty impressive!

Seems easy to defeat with frequency bounds checks.

But presumably there exist adversarial sounds that to a human seem like music, or gibberish, or a different phrase, but sound like "pay joe 1000 dollars" to the machine.

The sonic equivalent to these: https://arxiv.org/abs/1703.08603

Cool stuff! Question from your FAQ:

> Q: Will I be able to copy another person's voice?

> A: Yes but only if you have the authorization of the person whose voice is being copied.

Perhaps you can unpack that answer a bit? What's the authorization process?