So it's essentially a radar distance-following system with a very accurate (you hope) map.

This is just unbelievably lame.

A Tesla can do those things - even without the self-driving option - but can do it on ANY road.

Add in the self-driving option, and it'll stop at stop lights and stop signs, handle lane changing to pass slow vehicles and to get into the right lane for turning and to avoid roadworks, adjust it's speed if the adjoining lane is going slowly,

My Tesla drove 3,600 miles across the USA and back - pretty much all by itself.

These map-following/distance-maintaining systems are obsolete by at least a decade...time to start pointing this out and not keep telling people that they are amazing new futuristic things.

With SpaceX claiming that they'll shortly be able to launch 100 to 150 tonne payloads into orbit for $2 million - it's time for astronomers to build more space telescopes. With a 9m diameter cargo bay - and with most of the world's greatest telescopes having between 8 and 10m optics - it's clear that a fully capable, fully modern scope could be put into orbit for comparable construction cost to a similar device stuck on a mountain-top in the Atacarma desert (or whatever). They'll soon be able to do that without the need to fold the mirror (a'la James Webb) and without any compromises on mirror diameter. Furthermore, having high bandwidth satellite data streams from the very systems that the astronomers are complaining about - they'll be able to get better use from these machines. No light pollution, no seismic vibration issues, no dust on your optics, no atmospheric twinkling, etc, etc.

Just imagine a version of Hubble with a 9m mirror instead of it's present 2.4m...about ten times the light capturing ability.

Do that - and allow the very low Earth orbit to be used for the greater good of mankind.

This can easily become a win/win proposition...it's just a matter of the interested parties on both sides of the issue getting together to cooperate rather than fighting each other.

So the air is unbreathable and would kill you. The pressure will kill you. The concentrated acid rain will kill you. There's no food or water. The day/night lengths are unbearable - and that, combined with the cloud cover, makes solar power useless... ...and you're worried about a few immensely active volcanoes?

MYTH: Bitcoin is anonymous.

It certainly isn't - it's fairly easy to watch the IP addresses in the block chain - and you can (by design) follow the long chain of transactions leading to any particular money transfer.

Anyone with enough computer resources and the right software can determine who is paying who what.

We're pretty much a "cashless society" now - I haven't had actual cash in my wallet for more than two years. This is all very convenient - and drives online sales and so forth. But the GIGANTIC problem is that we're now (though higher prices) paying those credit card companies at least a few percent of everything we earn.

For doing what? Moving some numbers around inside a computer. That's an egregious amount of the economy - for doing something that's ridiculously easy.

So it makes sense to have the issuer of "money" (the treasury) handle the shuffling of numbers as a non-profit...and have people only use credit cards when they ACTUALLY need credit.

So it makes sense to have the government (or the treasury) set up a system of digital currency that's not providing "credit" but is cheap (or perhaps even free) for use when you're only shuffling money around.

This would also eliminate the business of cryptocurrency mining - money could be created in just the same was that coins and banknotes are...without all of the egregious amounts of power consumption, etc.

Done with care and with serious attention to security - this could be a very good thing.

So if you have your own solar panels - and you are "grid-tied" as most people are... Then your excess generating capacity is fed to the utility company to pay down your bills, So now you're a renewable energy provider...do you become liable to under this law?

And in other news: multiplying by ten is preferred over tenth-roots.

So...you grab your two kiddie car seats and one booster seat - and you call you RoboTaxi. You fart around for 10 minutes installing the car seats - you ride five minutes to the child care place - remove the car seats again - drop your kids off, then carry two car seats and a booster seat and call another RoboTaxi. Now you have to go to work...well, I guess you can stuff the car seats under your desk while you work...then, on the way home, more car seat installing and uninstalling. If you want to stop at the store and pick up diapers- where do you put the car seats while you go around the store?

Honestly - this can't work for a very large percentage of the population.

When you think about the REALITY of taking a cab everywhere - it can get VERY challenging.

In truth, the very concept of "source code" ceases to exist with AI. You have a blank AI - just a billion 0.5 weights...then you train the crap out of it and you have a billion random-looking numbers. That's it...that's all you have, There simply isn't any source code to share.

5 is already true for Cruise - it can only self-drive on US freeways & tollways.

9 is edging towards truth with Tesla where "One pedal driving" takes some getting used to - and "Cruise Control" is entirely different. Even worse, you sometimes need to re-learn after a software update!

We started off (I believe) with the "Space Beer" thing - it was kinda reasonable, it investigated how carbonation could be simulated - how taste bud react differently - that kind of thing.

Then there was the champagne bottle/glass that was some weird complicated device.

Then there was the "experiment" to see if cement would set in zero g (big surprise...it does!)

Then the specially built oven for baking cookies - (all of which were sent back to Earth, so the poor crew never got to try one!) ...and now this.

Admittedly Estee will have paid more for the privilege of splattering their names over the world's news feed than it costs to launch and retrieve their junk products and the time it takes the astronauts to snap a photo of them...but this really is the end.

If we seriously have no better reason to keep the ISS up there than this - then it's time to de-orbit it...actually it was time a couple of years ago.

It's told us everything we needed to know that it can tell us - and now it's done.

If the advertisers need this service then let's sell them the iSS for some bargain price. Maybe WPP, Omnicorp and Publicis could fund it.

Meanwhile - StarShip will have a larger internal pressurized space than the ISS, and doesn't need servicing missions. It'll be able to stay up there for at least 6 months (that's how long it'll take to get to Mars) - and we can just land it to resupply it.

De-orbit the ISS....this latest thing proves conclusively that it's over.

You are allowed to watch movies from tobacco companies - the restriction is on them. They aren't allowed to make them.

YouTube, Facebook, etc are allowed to decide for themselves that there are movies that they don't want to show.

Elon Musk has been talking for a couple of months about a "major rewrite" of the low level infrastructure of the autopilot code.

Previously, it had mostly been a 2D exercise - calculating a "safe driving" area in 2D and mapping obstacles into it from cameras and sensors.

Now it's going to be 3D from beginning to end.

But the important part of this is that many new features have been developed that depend on the new infrastructure - and pre-existing features are working much better when implemented on top of it.

The rewrite is supposed to be nearly complete - and when it is, all of those blocked features will become available.

This story (assumed to be true) would explain both the unpredicted slow progress over the past 6 months or so - and the claim for a sudden leap forward in the coming 6 months.

This explanation does also match the facts - there are videos of tests people have done. For example, if you take a traffic cone and lift it above your head - the Tesla doesn't see the cone as moved vertically - it sees it move rapidly off into the distance. This is because the AI thinks that traffic cones belong on the ground - and the only way to reconcile a cone up in the air - is to imagine it's far away and raised towards the horizon by perspective.

As a software engineer - I understand the upheavals that large scale refactoring can cause - and how features can get backed up waiting for the refactoring.

This claim from Mr Musk is quite credible - and it explains much of the facts as we know them.

LIDAR gives you range - and not much else.

The software to extract range from stereo cameras is pretty routine stuff these days - and with cameras that are moving at known speed and direction, it's even easier.

Since you can buy a whole bunch of cameras for $10 - I don't see the joy of LIDAR.

A Tesla car is likely to last 600,000 miles - and with next-gen batteries, for a million miles. An average gasoline car needs engine and/or transmission work that typically costs more than the car is worth at between 150,000 and 200,000 miles.

So unless you can show that a Tesla needs between 3 and 8 times as much resources to make as a regular car - this is a poor argument.

And that's even ignoring the CO2 savings over it's life. A gasoline engine is typically only 25% efficient - and the oil refinery only about 70% efficient. So overall, you're only using aout 20% of the energy from the crude oil to propel the car forward. Even the NASTIEST "brown coal" power station is 65 to 70% efficient - and the transmission grid, charging system and electric motor is about 90% efficient.

So, again, even with the NASTIEST, DIRTIEST electricity available - an electric car is about 3 times more efficient than a gasoline engined car.