Discontinuing the penny?

For once I want to endorse one of the Trump administration's proposals — discontinuing the penny (one-cent piece), which costs more than 3 cents to make.

But I think we should go farther, and discontinue all circulating coins except the quarter.

The half dollar is already effectively discontinued; it became unpopular back in 1964, when people started seeing it as a JFK/LBJ campaign piece. In the 1970s and 1980s, I never got a half dollar in change, even once.

The nickel costs appreciably more to make than its face value. But if you discontinue just the penny and nickel, you create an awkward situation where the quarter isn't the smallest coin but can't be broken into smaller coins.

More to the point, when the half-cent was discontinued in the 1850s, one cent had the purchasing power of about 35 cents today — that is, more than today's quarter. That was considered small enough.

People say that if we discontinue the penny, prices will go up. Only if people are shopping with pennies in the first place. Who does that any more? I haven't spent any cash at all in several weeks. For two or three years I have carried no coins in my pocket; there's paper money in my wallet, but it's very seldom accessed.

Federal taxes and some other large transactions are already rounded to the nearest whole dollar. And as a computer programmer, I would jump at the prospect of no longer handling fractions of dollars. Computers handle whole numbers much better than decimals. This is especially the case because 0.01 — that is, one cent — has no exact representation in binary, for the same reason that 1/3 does not have an exact representation in decimal. So to handle cents, the computer has to simulate decimal arithmetic in various roundabout ways. Handling whole numbers is built into computer hardware and goes much faster.

Some people may shed a tear that it's no longer 1958. But... it really is no longer 1958!

Permanent link to this entry

"Move fast and break things"

(From Facebook.)

At risk of having everyone hate me because I open my mouth about politics, I want to bring out an issue that has been lurking in the background.

I am in favor of trimming government spending and eliminating waste. But I do not consider rank-and-file federal employees, nor scientists and students supported by federal money, to be ENEMIES who need to be PUNISHED.

Apparently quite a few people do, and they feel a thrill at the prospect of making these people suddenly worse off. "The government doesn't owe anyone a job," said one person today — but when the government does hire someone, surely it has the same responsibilities as any other employer, to deal with them in good faith and not break promises or cause needless hardship.

"Move fast and break things" means "Move fast and hurt people" and is not good management, much as it thrills a certain kind of person. Waste can be trimmed in an orderly way. And most of the waste is in large programs, not small specialized ones.

Don't punish the employees and students. Give them an orderly path to a soft landing. In fact, look inside yourself. Ask whether you yourself want to make people worse off, and if so, why, and what it says about your character.

Addendum: Trump and Musk's campaign to "eliminate fraud and waste" strikes me as bizarre. It started with firing the Inspector Generals (whose job is to eliminate fraud and waste) and has shut down the Consumer Financial Protection Bureau. Public communications have been suspended so that the FDA cannot announce food recalls and the CDC cannot tell us their measurements of COVID and flu prevalence. Research grants and stipends have been cut off suddenly so that, as far as I can determine, graduate students and postdocs were suddenly left without their paychecks. This is not what good government looks like.

Permanent link to this entry

Solving a contradiction in the Bible

St. Paul became a Christian when he saw a blinding light and heard the voice of Jesus in Heaven. The people around him saw the light. It is unclear whether they heard the voice.

Acts 9:7 says, "The men who were traveling with him stood speechless, hearing the voice but seeing no one" (ESV).

Acts 22:9 says, "Now those who were with me saw the light but did not [hear] the voice of the one who was speaking to me" (ESV: "understand").

The context of the first verse seems to indicate that the light was visible to all, so that's not a contradiction. The question is what the men heard.

The Greek words for "voice" and "hear" are the same in both verses. There is a slight difference of grammar; the object of "hear" is in the genitive case in the first one and the accusative in the second. Scholars differ as to whether this indicates that in the second verse it meant "understand." That is not perfectly clear.

So how do we approach this?

Some would say, "See, this proves the Bible is all blarney." But I respond that even if this episode were total fiction, you'd expect a careful writer not to contradict himself, especially since the text was copied many, many times by scribes who would notice problems.

A second possibility is that — since two different people are describing the incident — one of them was mistaken or less informed. The first verse is the narrator of Acts (presumably St. Luke) speaking in his own voice; the second is Paul giving a speech.

But the third, the one for my money, is that the incident was something that could be honestly described both ways.

That is the approach I'd take to a contradiction in a book that didn't claim divine inspiration: Assume the author had something coherent in mind, and figure out what it was.

That brings us back to the interpretation that the men heard a sound (the word for "voice" can mean "sound") but did not understand the voice (the word for "hear" can mean "understand"). Case closed?

Permanent link to this entry

Another wide-field view of Monoceros

On February 3 I got the same equipment out again — Askar 200-mm f/4 telephoto lens, Altair 26C camera, and Losmandy GM811G mount with iOptron iGuider — and took aim at Monoceros again. Much time was wasted because the guide camera was loose in its holder, and this led to bad guiding, as well as error messages such as "Guiding stopped because the telescope has slewed" when it definitely hadn't.

But I ended up with a series of 75 2-minute exposures, which stacked very nicely. Here is a large part of the end result, followed by a more tightly cropped presentation of the central region. The blue-green color in some of the white areas is due to the Orion SkyGlow filter.

M42 (Orion Nebula) by Askar and Altair

Yesterday evening was clear and not too cold, and I decided to try the Altair 26C camera on my Askar ACL200 200-mm f/4 astrographic lens, a combination I had not yet tried. This is a stack of 28 2-minute exposures with an HDR (high dynamic range) effect applied to bring out detail in the bright areas.

Permanent link to this entry

A wide field in Monoceros

My second target for the night was part of the constellation Monoceros, including an almost uncatalogued nebula I've been paying attention to, plus many other dramatic clouds of dust and gas. Unfortunately, I forgot to cool the camera sensor, but the effect on picture quality was small. Like the previous picture, this was taken with the Askar 200-mm f/4 astrographic lens and Altair 26C camera, with an Orion SkyGlow broadband nebula filter. This one didn't have HDR effects applied, and unlike in the picture above, you're looking at the entire field of view. This is a stack of 49 2-minute exposures.

The state of AI

For two weeks, for several projects, I've been intensively surveying the current state of generative AI and LLMs. And here's something interesting...

I haven't heard even one clear mention of consciousness or "artificial general intelligence."

The objective now is to do useful work by mixing techniques any way that helps.

People are no longer trying to prove that a table of word frequencies can become conscious and take over the world. Instead, they mix LLMs with other software (both conventional AI and non-AI) to get things done.

Up-and-coming ideas include fine tuning (giving an LLM extra training on authoritative documents, to override less accurate information in the original training set), reinforcement learning (further training the LLM with feedback from human users and/or other AI, including good old-fashioned rule-based AI), agents (letting the generative system create a series of tasks rather than a series of words), and tools (letting the AI call upon other software to get exact answers).

This may be a bigger change than we had a couple of years ago when ChatGPT went public.

Permanent link to this entry

Batting near zero?

Now that we can track packages on the Internet, I'm seeing how much can go wrong.

A package that I mailed to Miami has reportedly been delivered twice, several days apart, at the same location. I mean they are reporting completed deliveries, not delivery attempts. I haven't heard that the recipient actually got it. It's parts from a broken laptop, but it's packaged like a normal laptop, and I suspect there's about to be a very disappointed thief in South Florida.

An auto part that I ordered from RockAuto in Las Vegas was due here yesterday, but instead, it's taking an unscheduled multi-day vacation in Tucson. It's a taillight cover, and I need it before driving the car in the rain, if possible.

And — this isn't about shipping, but rather telephone follies — one of Melody's prescriptions at a mail-order pharmacy was snagged for two weeks because they couldn't send a text message to our home telephone number. It's not a cell phone! It took them two weeks to get around to dialing it and talking. They got me on the phone, but if I hadn't been here, it would have made a recording. Yes, youngsters, there are still telephones in the world that don't receive texts.

That would be comical, but a two-week delay with a new prescription is not a good thing.

Permanent link to this entry