serioussecurity

This piece is a bit unsettling, most notably for it's being being barely coherent in structure, picking out seemingly random observations and facts.

The observations and facts all would fit nicely into an effort to cast Mann's intellectual mission in a negative light. Tying him to russia as the closing paragraph for no discernable reason? Suggesting his ideas were self-aggrandizing? Suggesting that he was trying to induce a system of false-belief through myth-matking? They're all observations thrown together without comment or an assertion of a common thread, and represent a remarkably cherry picked subsample from the book that supposedly forms the core of the article.

It feels like an opening salvo in an ad hominem attempt to dismantle a core figure in anti-Nazi thought? Coming from a journal intimately tied up with the neoconservative movement.

(Yeah, that's also an ad hominem, but the point about the article stands: it's barely coherent and the things it's suggesting form the start of an unsettling attempt to frame a narrative.)

Everyone here seems to think that if you have a good job you shouldn't be involved in labor organizing, because everyone else has it worse?

Isn't the point that everyone should be organizing because working conditions everywhere are much worse than they could be, and the organizing at Google is a case study for how organizing works at one particularly unusual company?

Density in this case referring to "kW per acre of solar panels", and the #s come out to something like covering all of Australia with solar panels, if I remember correctly.

An uptick in crop failures due to irregular growing seasons has lead to things like the Syrian civil war, which is a fairly direct cause of the immigrant crisis in Europe which is a fairly direct cause of the rise of ethno nationalism in Europe.

Despite high hopes, incompetent new comers discover they can't compete in new industry...

So many arm chair lawyers on HN. The parent is wrong. It's illegal because it violates the NLRA, not whistleblower protections.

"Protected concerted activity".

If you want a good primer, "Labor Law for the Rank and Filer" is a good one.

:-/ I make $300k / yr.

I live in a 10 x 18 room, in an apartment with 3 room mates that has an acceptable kitchen and a cramped living room.

I save > 50% of my income, which is nice.

But I also spend ~ 80k / yr on medical expenses.

I guess I'm upper class if "hoping to retire to a remote part of the world at around 35, and live off ~20k / yr til I die" is a reasonable hope. But also that's just ... ill.

I hate the "upper / lower / middle class" model because I don't think it really is a meaningful way of slicing society anymore? It's a very post-WW2 idea of how our society is organized and what's desirable.

I'd tend to start slicing along a set of lines like: -- urban / rural -- race -- age -- degree of community support / culture -- type of labor (knowledge, service, manual, ...)

I'm working on building long term community structures (one of my low-key goals is to buy and convert a church into a community space). Because the "I'm X class!!!" is dumb; it treats people as totally not dependent on their community. Unless you're willing to pay for community builders to do their work, you're not going to have any real community.

(btw, this is why I can't stand people skeptical of affordable housing and the like: why do you think it's pleasant to live in neighborhoods with a robust variety of people? because often people that aren't working all the time can do things like maintain a community garden. Which is only possible if they can live in your community.)

Most of the big tech companies (amazon, google, facebook most notably) have discovered an extremely strong correlation between latency and $$$. Every saved millisecond of average load time tends to generate an incredible amount of money (like, 10s of millions of dollars a year per millisecond).

Unsure if joking. Look up stuxnet

Their datasets aren't useful for the problems that make driving hard. Musk is making claims which are paper thin false. If he weren't making these kinds of claims I'd take him more seriously. I promise you that the reasons this is ludicrous are extremely solid.

Edit: fwiw I've actually taken a look at a number of waymos internal systems, and sat down with people who worked on autopilot.

This is one of those "you need to be an expert to evaluate the claim" problems.

Talk to anyone with research experience who are building world scale ML systems that are used by real people, and it's unambiguous that Elon Musk is a con man.

It comes down to small details, like the nature of the training data, the nature of the customizations, the nature of the business problems, the nature of the regulatory problems, the difference between highway driving and city driving. The enormous amount of special casing, the nature of the QA infrastructure, the capital on hand, the routing and mapping problems...

What Tesla has is a decent lane finding system. What Google has is an end to end system. It won't take them long to expand to all the major cities; solving the geofencing problem is easy, they already made street view.

And they were doing L4 successfully years ago. They're keeping drivers in the cars right now mostly because something going wrong is a PR risk, not because they need them. And L5 isn't a particularly important goal.

My guess is I'd have to write up 20 pages of detailed analysis at minimum to convince folks. In general if you don't have 5 to 10 years in depth experience in this space, you probably shouldn't consider qualified to have an opinion.

Because ad click through rates are so low and the signal is so incredibly sparse. It's all measured against a test set which the researchers rarely look at directly (because the data set is so large). So if something lifts ad vlicks by 1% and that's from 100 in 10k to 101 in 10k, they don't really care if 9k of the ads get appreciably dumber.

It's called "protected concerted activity". It is protected under the national labor relations act. These laws were passed in the context of mine owners hiring armed mecenaries (the pinkertons for example) to break up strikes. I highly recommend "labor law for the rank and filer" as a history and primer on the law here.

It's called "protected concerted activity". It is protected under the national labor relations act. These laws were passed in the context of mine owners hiring armed mecenaries (the pinkertons for example) to break up strikes.

The NLRA is pretty complicated and it comes down to what you can get the NLRB to agree about, but this does qualify as protected concerted activity. Anything which is collective organizing in the interests of the employees w.r.t. working conditions is protected. Protesting gender discrimination is literally a text book case. I recommend "labor law for the rank and filer" as a nice introduction to the field.

The history of labor is in this form of protest. Googlers are more comfortable than many others. The further down the chain you go the more precarious the workers and dependent on their pay checks, and the easier time management has retaliating.

This stuff is legally protected because before we had legal protections, conflicts devolved into straight out armed struggle, with employers hiring mercenaries (notably the pinkertons) to break up employee resistance with guns.

Yes they are. It's called "protected concerted activity" and it's protected under the national labor relations act.

These laws were passed because they were seen as better than the status quo, which was employers firebombing their employees or highering mecenaries to shoot them.

depends what you mean. If you think of it as a "replacement" of another launch they could have sold, then yes, it's the same cost. Lost profit opportunity is lost profit is a cost.

But if it means they can increase the # of rockets they build, they might pick up some efficiency gains on the back stretch. I'm pretty sure they're already struggling to keep up with demand, tho.

Strong disagree. I started learning machine learning back in 2010, before neural networks really took off. The understanding of where the field was before and why DNNs were useful advancement puts me way ahead of peers who don't understand what problems they're solving. I have an appreciation for the ways that those problems can still creep back into neural network research, despite them apparently being a black box to many researchers.

Maybe we reckon with the issue that "understand" doesn't mean anything in the absence of it's implications and we rephrase the question. What do you view as consequences of deciding that it "understands"?

Also selling it to the military to "increase lethality".

https://www.theverge.com/2019/2/22/18236116/microsoft-holole...

Book rec: The Body Keeps the Score.

Bessel van der Kolk has been one of the most important PTSD researchers of the last 40 years, and his work on the embodied aspects of trauma and stress is fantastic.

Things I found changed my life:

* read The Body Keeps the Score * get a massage at least once a month * Start a regular yoga practice * Start journalling with pen and paper * Find ways to get outside in nature / around water

God speed

Yoga and massage are great for chronic anxiety and depression.

Actually taking seriously the "your feelings live in your body, let yourself go where you feel" stuff while doing yoga / massage can be really emotionally cathartic.

Clearly you've never experienced a psychotic break....

uhhh, you mean "encrypted text messages were shared with Mueller's office"?

Yeah, to be clear, Watson failing is good news. This is marketing people losing their minds because they didn't understand the products they were selling, and hospitals failing to do due diligence.

I've talked to literally dozens of people / companies / academics running at problems in healthcare using deep learning and large data sets, and the recurrent theme is cowboys who pretty consistently fail to follow basic practices of experimental rigor. ML is frankly an area where more rigorous and clear regulation is desperately needed, because the potential to cut corners / cheat is so easy and the challenges of vetting these systems are specialized enough that most organizations won't successfully be able to catch fairly basic experimental mistakes.

Tesla are clowns. It's possible to do this right, but they're shipping a product they know isn't suited to the task.

(Lidar really does matter, but also extremely rigorous testing.)

So there's a couple things at play.

Lidar range is more than double visual range, in practice.

When figuring out collision avoidance paths for an object you actually end up approximating some np hard problems to find a path that won't have collisions and won't be too "careful".

This ends up being fairly computationally intensive, and adding the extra time significantly improves your planning. Doubling compute time tends to beat doubling your training dataset in terms of system quality, at this scale.

Extra time also turns a number of situations from "guaranteed kill" to "we can avoid the accident", because the car is traveling really fast and those extra seconds can be used to brake, find a new path, etc.

In visibility impaired situations, lidar and vision have different constraints and ways they fail, and the intersection of the two can significantly improve scene understanding ( see waymo's snow demo ).

In a lot of cases, path planning can be dramatically improved by having maps. If you're going into a curve and know the shape of the road, you can preload that and spend your time on more important tasks like object detection and path planning.

Etc etc etc.

This is absolutely not a domain for intuition and thought experiments. The pragmatics of the industry are highly intricate and responsive to constraints that are only visible if you've worked on this stuff.

I keep seeing this meme about Tesla and it keeps making me grumpy. Basic lane keeping is simple, and Tesla is only doing that. And they're doing it badly.

Tesla's autopilot is a PR move to try to raise more capital. They are not and will not be a serious player in driverless cars.

What they've got is a system doing object detection. Object detection is actually one of the easiest parts of driverless cars. Planning and interaction end up being much harder and requiring a lot of data, none of which Tesla has collected. (Also, in practice you need lidar and a fully mapped road, neither of which Tesla has; we're at least a decade from the top players who are actually collecting large scale urban / suburban data from dropping the map requirement, and there's really no reason not to use lidar as the price is dropping so quickly. Both reasons Tesla's "we've got lots of data" claim are BS...)