kahnjw

I'm certainly no expert but I think tiers of battery storage will be necessary. Some collocated with generation sites, some at distribution sites in cities, and probably local batteries in commercial real estate and neighborhoods if not households.

There are both local and global shifts in demand. If it were the case that only local storage options existed the system would be extremely inefficient during abnormally high peak demand that was localized to one area. Think of a heat spell in Central Valley of California. Storage would need to move from tiny batteries scattered throughout The Bay Area to be routed to the Central Valley to power AC units. It would be more efficient if there were larger local storage sites in Sacramento and Fresno.

On the other hand if we only had huge storage sites that are collocated with generation sites then as an individual you might be able to game the market through arbitrage with one big highly efficient battery. Take in power at night when its cheap and sell it when it's expensive during the day. Of course everyone will catch on to this and want a battery of their own. This depends on battery efficiency, but if you can get near the efficiency of the battery at the generation site then you can undercut their profit margin. You could at the very least use the local battery when possible to power your own home/business to save some money on your power bill.

Why is iOS a different story?

The opposite, it’s a massive target. I believe the reasoning is that because it is such a large target the black market doesn’t set the price. It gets set by the bounty price the maintainer offers. I’m not familiar with the dynamics of the zero-day black market so I wouldn’t know myself.

On the other hand zero days in OS kernels that run on millions of devices worldwide aren’t your ordinary black market exploits. I agree with publishing after 90 days. This incentivizes shipping patches quickly and fixing root causes later.

Allowing companies to take their time fixing critical security issues puts users at risk. That’s compelling enough for me. As an Apple user I’m a bit upset that they haven’t patched after over three months.

On the other hand some services from AWS are downright bad. For instance, Elastic Beanstalk is an operational nightmare. The one selling point is that it's easy to get started with, but from there it seems they rely on ops lock-in to pay the bills for this product. It's slow, clunky, and opaque. There are much better alternatives that don't force you to tradeoff usability and lock-in for initial ease of use.

Agreed. After looking at the a few videos it seems to go a bit beyond the basics, I don't think that means it's "advanced."

This got me thinking: what would an "advanced" Deep Learning/RL course even look like? The field is so immature it's hard to identify topics that are advanced. Maybe an advanced course could be a chronological paper survey? Certainly not a class that spends a lecture on optimization of a deep network.

I agree road and vehicle to vehicle comm is important, but the work being done now by these companies is comparatively much more important. If there is an implementation bug on either side of one of these integrations or if the communication medium is bogged down, then what?

These systems must be able to handle diverse and unpredictable situations without being able to communicate directly with other agents in the environment, and must be able to act safely relying solely on their own percepts.

Corporations use debt to inflate valuations in the short run, so that sounds correct.

https://fred.stlouisfed.org/series/TOTDTEUSQ163N

FAANG and most other tech valuations have grown enormously over the last 10 years, but they've roughly followed the growth of their revenue streams and potential monetization opportunities. There may be some short term corrections, but the notion that the majority of risk in the current economy is concentrated in FAANG and some tech IPOs seems overly simplistic and probably just wrong.

The GFC was a result of individuals taking on massive amounts of debt in aggregate. Corporate America has similarly taken on a lot of debt over the last decade for buy backs and growth. As the author of the article notes, the last 10 years have been great for cheap debt. Negative balances are still a risk and rates are going up. The companies taking on the most risk in terms of debt are for the most part not traditional tech. Think Chevron, GE, and other industrial and energy companies. I'd conclude that the majority of the economic risk is in fact not concentrated in FAANG or even most of tech, but elsewhere in the economy.

That doesn't mean some tech companies won't get caught in the storm. I think we could very easily see a couple mid-size players go the way of Yahoo (the Snaps and Ubers).

The author makes a good point, anonymizing data is hard. Unfortunately they don't mention differential privacy, a promising area of research that can help us solve these problems.

https://en.wikipedia.org/wiki/Differential_privacy

It seems to me that diffpriv is a nascent area of research that has not yet been bastardized by the business community. The complete opposite of a buzzword.

I agree with all of this, except are you really implementing bayesian models in SQL? I'm not sure that is what the author is advocating in the first place. There are tools for bayesian everything, I much prefer using those over hacking together some illegible SQL script.

Forecasting and prediction. Specific example: given a high quality dataset, forecasting sales is a good use case for machine learning. Of course, you could do this with SQL as well, but again, given a good dataset and a complex domain like sales, some relatively simple models will be hard to beat with a 10 line SQL query.

This seems like a strawman argument. The original reporting from the guardian didn’t call this a hack. Also, the most egregious actions taken by Facebook are similar to actions taken by other companies after large user data compromises (Uber, Equifax). Facebook/Zuckerberg made a decision not to inform users that their data was compromised by an application they, or their friends had used. That is negligence. That is why people should be upset with Facebook. That is why I have stopped using it. This has little to do with Facebook getting hacked, and much more to do with a lack of social responsibility on facebooks part.

I think you are missing her point. She isn't saying that all passionate people are affluent white males. My interpretation is that passion in a swe job description is code for affluent white male. I think she has a point, individuals who have time for passion projects, self study, or other endeavors are privileged. Many of them are white, male, and affluent.

Passion: strong and barely controllable emotion.

"sufficiently non-loathsome...motivate yourself to do it" is a far cry from passionate. Tangentially related @sarahmei on twitter wrote a thread about passion in tech. IMO she nails it: https://twitter.com/sarahmei/status/921366047122268165

I agree, I just don’t think it matters in the 3rd grade. Find the high performers in middle school and high school. If there are students that really need help earlier than that, or have learning disabilities they can either be pulled out, or given extra attention if possible, maybe a tutor. I’d lean towards a bit of extra attention. And yeah, it’s definitely a chip on my shoulder, it’s a motivator for me. When someone tells you you’re not smart, you can accept it or prove them wrong.

They'd do fine, because for the most part, the students who were placed in TAG math were the same students whose parents were on PTA or whose parents had the time and self entitlement to go to the school and not so politely request their child be placed in the program.

The notion that gifted students need a special environment to learn is comical. If they're actually talented and gifted, they should certainly be able to learn effectively in the same environment as "normal" students.

Learning to read and do arithmetic is important in the 3rd grade, but it isn't as important as learning social and communication skills. I don't deny that there are gifted individuals among us, I'm advocating that we wait on trying to identify the gifted ones at least until puberty.

I think it is fair, but I think there are ways to give students who need it extra attention without singling them out. Children are smarter than we give them credit for, they know when they're being judged or singled out by adults and peers.

>gifted kids do benefit from suitable gifted education

If you show me that other students don't benefit from suitable gifted education, then you'll have an argument. Claiming that treating children with gifted education increases their chance of success is obvious, but misses the point, does it not?

I can specifically remember back to when the TAG Math students (Talented and Gifted) were learning what negative numbers were. I kept hearing about them. I imagined what they might mean, how you use them. When I was finally introduced one year later, I was disappointed, you mean it's literally just subtraction? I could have done that one year ago. And same with all the other students in the normal math course, if they had the same level of attention and care.

The worst part was this: the students in TAG math had a student:teacher ratio of 10:1, while the normal math students had a ratio of 20:1. The remedial students had a ratio of 10:1, but the remedial teacher was from the special ed department, not a specialized math teacher that the TAG and normal students got. The TAG program at my school was praised, and the teacher acted as if she was doing God's work. Of course, she had children who were students at the school, and of course, they were in all of her classes.

Now, looking back on it, I realize how broken this frame of thinking is. Yes, students will benefit if you put them in an elite math course with more resources than other children. Other students will bare the cost. I think a better allocation of our time and money would be to spread resources evenly. If you really want to get ahead by one or two years, take summer classes in middle or high school. That way, the students bares the cost of the additional time and effort, and we can still evenly allocate resources so every eight year old gets the same opportunity.

EDITS: spelling bear -> bare

When I was in 3rd grade, I was placed in a remedial math course. My mother requested I be removed from it. I was. Today, I am a senior machine learning engineer working on my third math heavy degree.

Needless to say, I am also very skeptical of identifying gifted students at a young age. IMO society over-values so called natural talent, and trying to find that talent in an 8 year old is silly, with rare exception.

Now, many of my higher performing peers from 3rd grade are lost. Hopping from career to career, city to city without a sense of direction. A few of the high performers have gone on to do impressive things, but nothing shocking, nothing that is any more impressive than what my stoner friends and I were able to accomplish after a few extra years to mature.

It took me more time to find that what I really enjoyed was engineering and math. Of course, this is a first person account, so take it or leave it. Worth noting that there is a whole lot of very bad education research, so anyone whipping out journal articles to prove a point should be cautious.

The problem with treating "gifted" students differently is similar to gender bias in STEM, and racial bias in education overall. Some people are told they can, and other are implicitly told they can't. These messages get internalized by students, especially if you are unlike myself and don't have a mother who will stick up for you, and take you out of that remedial math course.

I don’t think op is claiming that method is tail recursive, just pointing out you can get away with using recursion and LRU cache.

From op:

>...avoiding the ire of the SEC is relatively easy if you set your mind to committing fraud without getting caught.

The experiment is a simple application of the economic theoretic framework to the statement above. Nothing more. There isn't all that much to argue about here, as long as you accept the framework upon which almost all of theoretical economics is based.

They are not fundamentally contradictory, everyone would know they are victims, still if it is in their net benefit to participate, they will. Again, this is an experiment with unrealistic assumptions (getting caught is an endogenous variable), it is intended to bring to bear the the absurdity of the statement that one can control their own fate if they commit fraud.

You can be sure exactly how easy it is because in a perfectly competitive economy everyone has perfect information, and the experiment's key assumption is getting away with fraud is a endogenous variable in the utility function. The fraudster can manipulate this variable to optimize their utility. This means that every actor knows exactly what they must do to avoid getting caught. It's a simplification so we can consider the implications of the notion that getting away with fraud is as easy as setting your mind to it. Another assumption is entities act solely in their own financial best interest. In this setup, everyone who will be better off by committing some form of fraud, which will be everyone who participates in the economy, will commit fraud.

I have asked him this. Paraphrasing his words: eventually, the SEC investigates virtually all securities frauds. Some cases are dropped if the act was criminal in nature and the FBI or law enforcement wants it, other cases are dropped if the suspect is deceased or fled, and of course there are cases where leads go cold and the case is back burnered.

The point is, fraud is risky, and it isn't easy. You basically have to be willing to give up your life as you know it. Sometimes I wonder whether fraudsters have an internal dialogue about whether it is worth the risk. And for the fraudsters who aren't psychopaths the guilt of defrauding friends and people who trust you must be enormous. Sounds like a stressful day job that I won't ever be entertaining.

Again, point is not just about SEC investigation. I'm pushing back against the notion that getting away with fraud is as simple as setting your mind to it.

Maybe you didn't carefully read my original post. I'm not claiming the SEC has a perfect record in finding frauds before the hurt anyone (that would be nearly impossible). I am pointing out that getting away with fraud is extremely difficult. Eventually, you'll get caught, the SEC will uncover everything, and you'll be broke forever. From op:

>...avoiding the ire of the SEC is relatively easy if you set your mind to committing fraud without getting caught.

This statement is mostly what I was responding to. Here is a thought experiment. For the sake of this experiment lets say our economy is perfectly competitive. If all an entity had to do to get away with fraud was "set your mind to not getting caught," then every entity that participated in the economy would be committing fraud because it is in their best interest. That sounds wrong, probably because it is. White collar criminals don't plan on getting caught, they do everything they can to get away with their crimes. To me, that qualifies as setting their mind to not getting caught. Well, they get caught anyway.

The companies you bring up (enron, worldcom, madoff) are outliers in the world of securities fraud. they achieved unprecedented levels of success and then failure. Using these cases are counterpoints to an argument that financial crimes are hard to get away with seems silly to me. After all, none of these frauds were successful in the end, instead, people ended up in prison.

What I'm hearing is "avoiding the ire of the SEC is relatively easy" because "Bernie Madoff."

IMO that is naive. He is an outlier. Sort of like saying committing mass acts of terrorism is easy because 911.