From PBS.

Billion-dollar weather and climate disasters broke U.S. record in 2023, NOAA says

I was introduced to computers during the microprocessor / home computer era of the late 1970s and early 1980s. It was a bunch of tinkerers and amateurs who went on to found companies like Apple and Microsoft. When I turned pro and went to school to study computing my influences were a group of older European academics like Hoare, Dijkstra and Wirth. These were the folks who really lead the way and solved the problems that nobody else could solve. We wouldn’t be where we are without them.

REMEMBERING NIKLAUS WIRTH: FATHER OF PASCAL AND INSPIRATION TO MANY

I happen to have inherited a Purple Heart. It was my grandmother’s brother, Anthony Scaccia, who died at the end of WW II. A fascinating read from the History News Network.

75 Years Later, Purple Hearts Made for an Invasion of Japan are Still Being Awarded

Was wondering how many solar panels it would take to power the US. Seems the US produces about 4,000 teraWatt-hours of electricity per year. That is 4 x 10e15. A modern solar panel will produce about 1800 Watt-hours per day or 657,000 Watt-hours per year. That’s 6.57 x 10e5. Dividing the production by production per panel gives the number of panels, or 1.6 x 10e10 panels. That 16 billion panels. Roughly 50 panels per person. This is for all electricity, including industrial uses. Sounds about right.

Panels are about 15 square feet each which gives 240 billion square feet. There are about 28 million square feet in a square mile, so we are looking at a bit over 8,000 square miles, an area about the size of New Jersey. Again, sounds about right. Thats about 5% of the US covered in solar panels. To be fair, the US is a big place with lots of open area. And nobody thinks 100% solar is a near term solution. But looking at the numbers, this isn’t far fetched.

The ice maker in our somewhat new Korean manufactured refrigerator is broken. A quick Google reveals that it is a common problem and a class action lawsuit may be in the works. My wife decided to contact The Manufacturer and see if they would fix it without charge. Texting on-line she was able to get a confirmation from the Manufacturer that yes, we could get it fixed, for free.

When my wife went to schedule a repair, she couldn’t get the charges waived. Around this time she gets an email to rate the response. One of the first questions was: were you aware that this was a chatBot handling your service? Uh, no. But that makes sense.

So AI promised us a free repair, but there seems to be no record of this, except the log of the chat. Will the Manufacturer honor the promises of its AI? We will see.

I was fortunate to meet Eugenio at the very end of my career. He was doing some fascinating AI work, mostly in vision. I ran across his Medium pages today. Great stuff from a guy who has a no-nonsense approach and a good grip on both the theoretical and practical aspects of the technology.

Eugenio Culurciello

From Roy Longbottom.

“In 1978, the Cray 1 supercomputer cost $7 Million, weighed 10,500 pounds and had a 115 kilowatt power supply. It was, by far, the fastest computer in the world. The Raspberry Pi costs around $70 (CPU board, case, power supply, SD card), weighs a few ounces, uses a 5 watt power supply and is more than 4.5 times faster than the Cray 1”

Cray 1 Supercomputer Performance Comparisons With Home Computers Phones and Tablets

From Kent Hendricks.

52 things I learned in 2023

I am a long time crypto skeptic. Not that I don’t think the underlying technology does what it claims to do. It’s just that as an engineer (by vocation and avocation) I just don’t see a use case for cryptocurrencies other than money laundering.

A key piece of money laundering involves “on ramps” and “off ramps”. How to get the money into crypto, then how to get it back out and into the regulated banking system. Banks, or probably more correctly banking regulators, don’t want money from anonymous sources sloshing into and out of banks, for all the usual reasons.

I recently read Michael Lewis’s book Going Infinite about Sam Bankman-Fried and the collapse of FTX, mostly because I like Michael Lewis and though he could clear up the murky financial situation at FTX for me. I recommend the book but kept thinking of one very small comment. He implies that the recent failure of Silicon Valley Bank (SVB) was tied to crypto. I was, and still am intrigued.

Googling anything about crypto is all but useless, but I did come to the conclusion that there was something of a run on SVB to cover crypto losses on other platforms. This makes sense. It’s hard to blame cryptocurrencies directly, and it’s hard to imagine what to do about such a situation.

it also occurred to me that FTX was using its internal hedge fund (slush fund?) Alameda as a stealth on ramp / off ramp for their crypto products. This was in a way similar to how Silicon Valley Crypto Bros were using SVB as a sort of stealth on ramp / off ramp into the legitimate banking system. When crypto crashed there was a run on the regulated part of the banking system that provided the ramps to prop up the crypto losses. There is lots more to all of this, but I am beginning to understand the problems of linking unregulated cryptocurrency to the regulated banking system.

Major crypto coins stabilise after U.S. intervenes on SVB collapse

Now I understand what Elon Musk is doing. He is building an army of killer robots. From the Daily Mail.

Tesla robot ATTACKS an engineer at company’s Texas factory during violent malfunction – leaving ‘trail of blood’ and forcing workers to hit emergency shutdown button