guimarin

A very good heuristic is to compare their total $ costs for purchase and use, and then think about what sorts of externalities that doesn't cover. I've found over and over again that $ cost pretty much equals energy and therefore pollution cost, less the externalities. But the externalities are really super important, and of course, engine efficiencies and where in the cycle the energy is liberated from it's chemical bonds is important.

Child has it. The way this study is written, we're asked to compare more or less the carbon impact of an empty gas tank and the battery pack. That seems a bit unrealistic to me. yes the drivetrain is different and therefore more costly, but when it comes down to it, looking at the examples in the paper, the comparison is between the battery pack and an empty gas tank. That seems disingenuous to the broader point of carbon emissions for the entire life-cycle of the vehicles, which presumably would include use.

that study is problematic because while they look at the cradle to gate cost of the electric drivertrain including battery, comparing to the ICE car drivetrain, they don't look at all the cost of Petrol + all the unaccounted for externalities that go into it's extraction and delivery. A good comparison would include the cost of electricity extraction both for the ICE car and the electric and then do the comparison. As it stands, this study tells us that Electric cars are more 'expensive' than ICE cars.

This is true, but centralized power generation is a lot more efficient then creating the power individually at the point of consumption like ICE cars do today. I believe that an ICE motor is something like 30% fuel efficient, Electric something like 80-90% and a combined cycle blah blah gas turbine is more like 60-70%. So no, it won't be the full 14% but it won't be a few points either, probably safe to say 5-10% reduction.

This is a follow-up from their announcement a year or so ago that the two would combine forces to create an open sourced framework/playground for AI learning. This announcement is simply that promise made publicly manifest.

This seems like such a small thing right now. Looking forward to the interconnectedness of people and machines, this is going to be a watershed event for a lot of people. Cochlear implants are in 10s of thousands of people today, and being able to directly connect them and other hearing aids to mobile supercomputers is a huge huge step. I cannot even imagine what new interaction modalities emerge as a result of this.

I have a friend at Intel who works in the AR division and their comment to me is that in one way or another all AR/MR glasses are running Intel silicon. I asked them specifically, does that include, Hololens, Apple's Project, and a couple others specifically, and they answered in the affirmative. YMMV though.

Altera, an Intel subsidiary. [1]

1. https://www.bloomberg.com/news/articles/2017-07-24/quest-for...

My dad was not a successful founder. He quit his job at 40 or so to start a company. He failed miserably, and despite us being fairly well off, went into a downward spiral. He is divorced, alone, and living in a trailer park. Every time I visit him in his doublewide in Campbell, still same furniture as person who died in it had, I expect him to be hanging from the rafters. He was a serial startup employee with over 20 years tech startup experience before starting. His startup destroyed his life. the YC, anyone can become a founder mythology is toxic. My dad was rich from previous exits, had everything going for him, and had a good idea/execution. He just couldn't boil the ocean and it destroyed him.

This is the craziest thing I've written in 2017 but: I sincerely believe we will have consumer fusion reactors by 2030 if not sooner. The work of MIT SPARC combined with graphene manufacture will make this a reality. Combine this with all the investment in battery load leveling and electric vehicles and you'll see a very rapid energy cost reduction and deployment. Probably 10 years once the design is functioning. As to the boats, these reactors are well sized for boats. My guess is ICE Air-travel will be completely eliminated for mass consumption by 2070 if not 2050, replaced by faster cheaper evacuated trains and electric short-range aircraft. The power density of li-ion batteries at roughly 350WHr/liter is not great, I'd wager we can easily get to 700WHr/liter with graphene supercapacitors utilizing a combination of 3D layering and fractal interaction. I've seen an experimental graphene supercap with fractal interaction at 1200Whr/Liter but it had a tendency to explode with only a few cycles. The normal graphene supercaps go like 100k cycles with no degradation, or 30 times longer than lithium ion cobalt (what tesla uses). We are so freaking close to cheap graphene it's ridiculous. That's the technology no one sees coming, it's just about to enter the productivity plateau. /end crazy person talk.

10% -> 2025; 20% -> 2030; 50% -> 2040; 95% -> 2050. India/Africa are included there. I think 95% of potential GHG emmissions (not just indexed to today but to 'present time') will be crossed in like 2060 or so. Renewables are going extremely quickly. we are in the steep part of the s-curve and there is no reason to believe we wont' get a few doublings due to battery cost decreases and deployed solar cost efficiencies. The transition will be a lot faster than ppl believe.

for the most part, I don't consider commodity equipment to be 'expensive' even if the price point is expensive. I've observed that there are usually two classes of products in any market, the commodity bottom 80% and then the premium top 20%. My comment is about products that inhabit this 20%. To date, and from my point of view, there is no product running android which falls in this premium category. With the Nexus 7 tablet specifically, I see a commodity android tablet with slightly better build quality and some more expensive components. Not a premium product like the iPad Pro.

Textbooks? Seriously though I think there is a movement among the rich in America/world to buy products that are simple to use, high quality, and durable. The problem in my view is not that these products don't exist but that it's very hard to find them and verify quality in production over time.

We should not be employing people in jobs that can be automated. We should try and automate everything. We should develop technologies, processes, and abilities so that everyone can learn new things. The people that want to learn will be incredibly leveraged and provide a ton of value to society. Those who don't want to learn should be given enough for basic subsistence. A stipend which covers food, shelter, clothing, catastrophic healthcare, reasonable water access, and unlimited data. I would also push that all humans can be close to nature in some way, be it a park or otherwise. We have the technology to do this. Instead we have protectionism and fear.

One easy way around this is to do what all the NYC foundations do. That is, give to other people's charities, and they in turn will give to yours. That is only small potatoes compared to donating to a political campaign and taking the money you don't spend on the campaign off the table tax free. A lot more things make sense through that lens. I could imagine a situation where Google who doesn't benefit from every OSS project, gives to project that Facebook benefits from, and Facebook to one's that Google does ad infinitem.

1. Seems like a business opportunity.

2. I believe we're talking about network accessible devices here. Otherwise why would parent make comment about air-gapping infeasible?

seems like a false choice. why couldn't i understand what functions these machines exactly needed and then put a small linux box between them and the network which monitors all incoming/outgoing traffic and only allows allow-list items to pass on to the ancient XP device.

You've just described buildzoom.com exactly. Check them out.

Pinterest, Houzz, and build.com are all really interesting points on the same spectrum. Pinterest currently doesn't do a great job of monetizing it's various communities, but is the leader in terms of customer volume and is getting better at delivering advertising value to both its customers and advertisers. build.com started on the opposite side by creating the relationships with all the direct suppliers/middlemen, and then giving people a chance to create deal directly with them. Houzz is a great combination of both. I would call it a specific vertical unbundling of Pinterest with the supplier relationships of build.com.

It will be interesting to see Pinterest build these direct relationships with sellers rather than build their own affiliate or else advertising platform. The big open for me, but probably not investors valuing them at 10B+ is will they be able to capture all that purchase intent themselves, or will they ossify like craigslist unable to deliver on any of their vertical specific values while slowing being cut up by competitors like the Knot, Houzz, and others. From what I read in their press releases and articles about the company, the Pinterest team seems miles ahead of craigslist at the same point in their development.

The interesting dynamic here for me is what happens as purchase intent advertising shifts from Google and web-search to Pinterest and others. Seems obvious that pictures are more accessible and valuable than text but I'm sure there is an army of statisticians at Google who can argue otherwise with me.

Also, it's really validating to see Houzz raise so much money at such a good valuation. The team there has been working hard in the shadow of the bigger companies and has created a real competitor in their vertical. They are here to stay.

You're absolutely right. Totally missed that nuance (damn word ordering). It does change my point to be a bit less strong, but I guess I'd say that I'm still happy that the number is large in the rest of world. This is a global problem.

It's interesting to me that Norway is such a leader in deployed EVs simply because of their Sovereign Wealth Fund getting so much of its revenues from Oil interest. That would be like Saudi Arabia having 30% EVs, seems crazy.

The other fun number is that the US was only 160k out of the 2M vehicles. Much more in line with the US population proportion to the world. A signal perhaps of the US's waining global leadership. Perhaps in the next ten years America will shrink back into a regional superpower.

Is there a case of an iPhone or iPad combusting like this? I haven't heard one. Not to say that there isn't but I'm more likely to believe it was a Samsung device than an Apple one simply because the former actually has a lot of examples of combustion.

As a Princeton Alumnus whose application data might actually be released in this probe[if they go back a lot of years], I have the following thought. I think people put way to much weight/stock in standardized test scores. I scored very highly in both GPA and PSAT/SAT, etc tests in my quest to get into an Ivy League school. Looking back, I think my high test scores were more a measure of my culture and socioeconomic background, then my aptitude. A lot of research seems to back up that inference as well. The problem for admissions officers is that they need an objective standard to create something objective, and they probably fail a lot at it. Thankfully people seem to be fairly adaptable and more or less people figure out their way in life. And also thankfully, Admissions people treat their job very seriously and try to 'craft' classes that they think would benefit all. Is it 100% accurate, probably not. But I don't think anyone reading this forum would actually like a Judicial System that is 100% decided by a dispassionate computer either. For me the two camps are analogous. Justice by computer, and school admission by test score. Human systems are a lot more nuanced than that in my experience. Maybe it works for CalTech, but every school is not CalTech and neither is every undergraduate a CalTech 'student'.

Underlying this whole debate is the idea of school = earnings potential. I think it's a false dichotomy to say that a great school = high earnings for life. They are definitely correlated, but in my experience, my classmates who have graduated and worked really hard are successful, and those that haven't by and large are not. I think if you took those same successful people and put them into different circumstance, re school, that they are the type of people to overcome that 'perceived' disadvantage. There is a weighting of privilege, but my lived experience indicates that it's more socio-economic and cultural than school based. Even at Princeton, the rich kids had their cliques that have continued into adulthood. So like everything, once at Princeton it was, what elite more selective group can I join to get ahead. I think this is very toxic thinking, though unfortunately intellectually easy so a lot of people do it. Figure out what matters to you and focus on that. The people I've seen do that, seem to find their way.

Youth is wasted on the young.

This might be a case where no one has actually tried it with a maker mentality. Musk was saying the other day that you need 1 atmosphere of pressure capability to create a vacuum, but 5-6 to hold back the watertable related to his tunneling network. Seems like from a first principles perspective making a vacuum should be much easier than an underwater tunnel. I'm guessing that you could create a partial vacuum in such a tunnel fairly easily. But I haven't 'done it' and that's the rub.

good argument that all internet businesses were invented in the late 90s bubble. nearly all were too early.

Or of course, things could take an altogether different turn. I remember when GPS was introduced in 96 for civilians and no one could have predicted Uber at that time, though all the pieces were there. BE does a great job of synthesizing possible directions from what's available to him in albeit a really good position at a16z. However, there are many people out there quietly building technology that changes the game completely. Making affordable graphene is one such unknown at this point.

I'm guessing they won't move very far up the stack for the time being. It seems to me that once a process is able to be software-itized at least two things happen. One it comes under that old adage that any computing program/application can be done for half the resources/time every 18 months. The second is that producer surplus (profit) that emerges from an information asymmetry becomes consumer surplus when that asymmetry is able to be defined in software. All supply side economies of scale seem to be disappearing and what we're left with are companies that have a monopoly on some raw input or a demand side economy of scale (network effect).

Do you have a source/ranking I can take a look at here? Hadn't heard of that and it totally changes the game.(no pun intended)

To answer your question is complicated because of the existence of Centaurs[1]. I will say that the fact that entities like Renaissance [2] exist is proof that there are still some humans that are better than machine learning algorithms. Now whether Renaissance employs Centaurs is hard to tell. My statement though is that machines reduce the number of experts because they reduce the population looking at a problem and my suspicion is that experts are uniformly distributed. So in a population of 100, there might be 1 expert in a population of 10000, 100 experts and 1 super expert, etc.

I chose the financial markets because you could say that it is the perfect competitive space for this kind of evaluation, though I wouldn't be surprised if Centaurs or humans existed in other areas as well which are better than machine learning algorithms. Still the point stands that for people in the 91st to 98th percentile a lot of value will probably be lost when humans start going away from practices in droves. Another acceleration for the 1%.

1. https://en.wikipedia.org/wiki/Advanced_Chess 2. https://en.wikipedia.org/wiki/Renaissance_Technologies