Matt's Homepage

Excellent article and something with extraordinary repercussions.

The world’s largest supplier of mothersboards, Supermicro, was caught adding spy chips to silicon they fabed for such companies as Amazon and Apple. All these parts had one thing in common: a Chinese contractor. One of the first cases was caught at Portland’s own Elemental technologies. A 3 year investigation showed almost 30 companies had their hardware infiltrated.

What does the chip do? Basically – it opens everything.

This system could let the attackers alter how the device functioned, line by line, however they wanted, leaving no one the wiser. To understand the power that would give them, take this hypothetical example: Somewhere in the Linux operating system, which runs in many servers, is code that authorizes a user by verifying a typed password against a stored encrypted one. An implanted chip can alter part of that code so the server won’t check for a password—and presto! A secure machine is open to any and all users. A chip can also steal encryption keys for secure communications, block security updates that would neutralize the attack, and open up new pathways to the internet. Should some anomaly be noticed, it would likely be cast as an unexplained oddity.

The strength of the Earth’s magnetic field is about 30 microtesla. The magnets in an MRI machine clock in at about 3 tesla, and the approximate magnetic field of a white dwarf star is about 100 tesla.

So just think about how powerful this 1,200-tesla magnet created by Shojiro Takeyama and his colleagues at the Institute for Solid State Physics at the University of Tokyo.

To achieve that intensity, Takeyama and his team pump megajoules of energy into a small, precisely engineered electromagnetic coil, the inner lining of which then collapses on itself at Mach 15 — that’s more than 3 miles per second (5 kilometers per second). As it collapses, the magnetic field inside gets squeezed into a tighter and tighter space, until its force peaks at a tesla reading unimaginable in conventional magnets. Fragments of a second later, the coil collapses entirely, destroying itself.

The last time Takeyama switched on his super-strong magnet, it blew out the heavy door of the lab that contained the machinery. Check it out here:

Everyone has seen those phone apps that make your photos look like the style of a famous painting. It is often called transferring artistic style. Code is easily available for you to write one yourself. Freshman in Stanford CS 231n do this today using neural nets.

Researchers have now extended that technique to video using some interesting new optical flow techniques. Will we even need animators in the future?

Autonomous cars are getting all the press, but there is an even more disruptive side to self-driving vehicles that will almost certainly come first – autonomous farming.

Imagine running a farm completely from your study? Sending fleets of tractors and harvesters to work 24-7 without a single human setting foot in the field. They can be timed to plant, harvest, or plow when conditions are optimal. Be monitored remotely by camera and even be driven remotely.

They could be combined with small drones or robots that are able to do fine labor – like weeding without damaging the plants and using a minimal amount of chemicals:

This isn’t just about reducing labor efforts – it’s actually a potentially huge jump in productivity and capital outlay as well. Imagine fields that can be analyzed and automatically planted based on market conditions, soil conditions – all to maximize profit and production.

It could radically reduce costs and environmental impact of chemicals used to feed, weed, and protect by insecticides. Imagine a machine that could drive over a field – targetting weeds and plants and give it exactly the right amount of insecticides, feed, and weed killer on an individual level.

This technology is not science fiction – it’s here in prototypes now. It should become ubiquitous in the next decade or two. Here’s a good overview of what’s coming and already in development

More proof that seeing is NOT believing.

I’ve posted numerous research papers that show how easy it is to make videos do what you’d like. How to take snippets of voices and make people say whatever you’d like. We can even now make people do dances on video they can’t do in real life. How easy is this to do? Deep fakes are shockingly easy as it turns out and only needs a few savy folks to make one.

Here’s the latest example where former US president Obama is made to say whatever the actor behind them wants.