Posted on
by

I’m pretty down on AI. I view Large Language Models as a very expensive way to generate strings of words. Other things that were called AI when they were purely speculative (for example, chess-playing computers), quit being called AI once the thing actually existed.

An open notebook and a pen, with a clock behind them.

I know there are other people who find AI useful as a way to minimize the amount of work they need to do, although again I’m very dubious—as soon as you let AI do something for you without double-checking everything it does, it’s going to screw up badly. (That is, you are going to to use an expensive, environmentally destructive, copyright-thieving tool to screw yourself up badly.)

It’s from that perspective that I can see this odd, narrow use of an LLM perhaps useful. (I actually got this idea from a fitness YouTuber, but he said he was going to delete his video, so I don’t see much value in linking to it.)

His suggestion was to use an LLM as a journaling tool. Each day, in the morning or in the evening, dictate a journal entry to an LLM. Do a kind of a brain dump of all the things that are worrying you or exciting you. List the things you need to do and the things you’re expecting other people to do. List your good ideas. List your long-term plans and your progress on your long-term plans. List your successes, your failures, your anxieties, interesting facts you came upon during the day, and so on. Then, once you’ve finished your brain dump, ask the LLM to organize those lists: Give me a list of things to do today. Give me a list of blind spots I haven’t been thinking of. Suggest a plan of action for addressing my issues. Tell me if there’s any easy way to solve multiple problems with a single action.

Now, I think it’s very unlikely that an LLM is going to come up with anything genuinely insightful in response to these prompts. But here’s the thing: Your journal isn’t going to either. The value of journaling is that you’re regularly thinking about this stuff, and you’re giving yourself a chance to deal with your stresses in a compartmented way that makes them less likely to spill over into areas of your life where they’re more likely to be harmful.

The things LLMs do—like generating summaries, making lists, expanding on ideas, suggesting alternatives—are really risky in any use case where getting it right matters. We all know the stories of people who used an LLM to write a legal brief, and the LLM hallucinated citations to cases, complete with footnotes, that were all lies. But I think in this case, none of those things matter much.

You aren’t required to have footnotes in your journal, or even to tell the truth. If you ask your LLM-supported journal to suggest blind spots that you’ve missed, and its suggestions are all either obvious or completely off-point, it costs you nothing but the 10 seconds you waste reading the list. If you say, “What are the top five things I need to get done today?” and it gives you a list that doesn’t include the very most important thing, you’re probably going to notice and get that thing done anyway.

Without having tried it, I can imagine that this might be a useful tool, or at least a harmless one. If all having the LLM in the background does is provide a bit of novelty that gets you started journaling again, even just that seems worthwhile.

One risk to keep in mind: If you’re telling an LLM things that are confidential—business secrets, personal secrets, other people’s secrets—you can’t have any confidence that the LLM isn’t going to feed all that info right into its training data and spit it out to some other LLM user. But pretty soon we’ll all be able to train up our own personal journaling LLM that doesn’t share its info with others, so I don’t think that’s a long-term problem.

Of course, I’m eliding the ethical problem of LLMs having been trained on stolen texts, but I don’t think that’s something the users are obliged to try to fix on their own. Rather, the creators of LLMs should be forced to turn their revenues over to the copyright holders of the people whose texts were stolen.

If you’ve been journaling with an LLM, I’d be interested to hear about it. And if you know of an adequately powered LLM that can be run on a personal computer without sharing any data off the machine, I’d be interested in hearing about that too.

Posted on
by

I wanted a workout to practice my Meyer fencing stance—a workout more interesting than just standing in the stance for a minute or two.

Mark Wildman, in one of his a live Q&A videos, suggested the mace drop swing as a useful exercise for someone doing longsword. (It was in response to a question I asked about improving arm strength and endurance for holding your arms forward and overhead at full extension for extended periods, as one does in longsword.) He had suggested doing it in Warrior 2, but specifically mentioned that you could do it in whatever stance went with your longsword style; it just got harder as your stance got wider.

Besides being boring, just standing in a Meyer stance for a minute or two seemed like a missed opportunity; even a modest challenge to your stability in the stance seemed like it might pay off in strength, flexibility, and control of your stance.

So here’s the workout I came up with:

Get in your best Meyer stance, with your mace in your front hand. Execute 5 drop swings, checking your stance after each rep. Shift the mace to your rear hand and repeat. Take one passing step forward. (Your mace will now be in your front hand.) Repeat five swings with the front hand and five swings with the back hand.

That’s one set.

Here it is as a video:

My plan is to gradually add sets until I can comfortably stay in the Meyer stance for 5 or 10 minutes, to build the habit and capability of keeping a good stance while it is challenged by a shifting weight.

I think my stance is okay here. Of course, there isn’t just one Meyer stance. This image from Meyer’s treatise show the range pretty well:

Plate K from Meyer's treatise, showing fencers in various versions of a Meyer stance
Source: https://wiktenauer.com/index.php?curid=43906

The front two figures are both in what I think of as a basic Meyer stance. The two figures behind them are also in Meyer stances, the one on the right in something of a lunge, the one on the left in a more upright stance.

My drop swings clearly need a lot of work (do not copy mine!), but that basically comes along for free as I do the stance workouts. (I wrote a post called Fitness training for longsword, Mark Wildman style that embeds two Mark Wildman videos of the Warrior 2 stance mace drop swing, if you want to see someone doing it better.)

I’m doing the swings with my 5 lb mace. I have a 10 lb mace that I’ll want to move up to, once I have the drop swings a bit more under control.

Posted on
by

Some time ago I made a mistake. Ashley was out on the patio, and didn’t want to come in, but I needed to leave. So I bribed her back in with a treat. This turned her into a monster. Knowing that she could get a treat for coming back in, she started resisting ever coming back in without a treat.

A black dog about to come back in a sliding glass door

Breaking her of that behavior by just not giving her treats to come back in would probably have worked eventually, but it seemed like a lot of difficult work.

So I came up with an alternative that has worked pretty well.

What I do now is never bribe her back in with a treat, but if she comes in when I tell her to—if she does what I want without resistance—once a day I’ll reward her with a Greenie (one of her favorite treats). (My wife does the same, so Ashley gets two Greenies most days.)

This puts us very much on the right side of behavioral conditioning. Since it’s an intermittent reward, it’s powerfully effective. And as long as we’re pretty good about only giving her the treat when she does exactly the right thing, she’ll start doing exactly the right thing every time, in hopes of getting the rare reward.

As I say, it has worked pretty well. There are a couple of things that aren’t quite perfect.

First, since she has learned that what she needs to do is go out and come back in again, she has started asking to be let out and then coming right back in again, hoping for a treat. We can deal with that simply by not giving her a treat when that’s what she’s done.

Second, and this is more a source of amusement than an actual problem, when the weather is bad and she doesn’t want to go out, she’ll go to the door, wait for me to open it, refuse to go out, wait for me to close the door, and then sit down to get her treat. I can totally understand her logic here: Isn’t just not going out the same as going out and coming right back in again?

She’s a good dog.

Anyway, shared in case other people are trying to cure their dog of demanding a bribe to do the right thing. Wait for your dog to “do the right thing” on its own, then give it a treat very occasionally. The dog will catch on pretty quickly.

Posted on
by

We have benefited enormously from the vast economies of scale in the vaccine industry. Because childhood vaccines were mandated, the companies that made them could be confident that they’d be able to sell large numbers. That made it worth both doing the research and investing in capacity.

Even flu vaccines have benefited, because government agencies got a bunch of scientists to come together and produce their best guess as to what strains to vaccinate against each year, so that there only had to be one vaccine that everyone got, and mandating that insurance companies had to pay for it.

But with the current administration in the U.S. suggesting that vaccines are generally bad, I fear we’re going to see less of that: fewer mandates are going to mean fewer vaccines being administered. Obviously that’s going to mean more sick people, which is really bad. But almost as bad, it’s going to reduce the economies of scale, meaning that the per-shot cost of vaccines are likely to rise significantly.

This all got me to thinking, what would a post-mandated-vaccines world look like?

Well, only smart people would buy vaccines, and only rich people would be able to afford them.

How many people are both smart and rich? And how rich would you have to be? Depending on how much prices went up, maybe only the top 50% would be able to afford them, maybe the top 10%, maybe the top 1%.

One small upside might be that boutique vaccine shops could find it worthwhile to make better vaccines—modestly better effectiveness, modestly reduced side-effects—because there’d be vaccine competition.

Really, though, there was always a strong push for that stuff for mandated vaccines, because if you’re going to give out 300 million doses, even a tiny improvement is going to really matter.

Still, I read a year or so ago about a version of the Covid-19 vaccine that produced much longer-lasting immunity being discontinued because they couldn’t sell enough of it, because it wasn’t mandated. That’s the sort of thing that might get better in a post-mandate world.

Won’t be a net win for society. Probably not even a net win for the 1% who can afford whatever bespoke vaccines they want, because it costs billions to research and test a vaccine, and even the 1% can’t afford that, unless they all get together and fund joint projects.

A selfie showing the bandage on my arm where I just got my Covid shot

These thoughts brought to you by me getting my Covid shot now rather than waiting until just a few weeks before I go visit my 92-year-old mother—because who knows if it’ll be available then?

Posted on
by

On Jackie’s last visit to see our friend Rosie, they visited a Navajo trading post, and Jackie bought a Navajo blanket. This morning we got it hung up in the living room.

A Navajo blanket hanging on the wall between a print of Bluebirds and a print of a White Buffalo Calf

This is directly over my living room chair, a spot where I’ve long resisted hanging framed pictures, out of an irrational fear that they might fall on my head. With a small blanket, I figure it’s no big deal.

Posted on
by

Early in the pandemic I bought a set of gymnastic rings, started training with them, and got great results. They have a few downsides, though: Since I hang them up outdoors it’s no fun to us them when its cold or wet or windy. Also, the easiest place to hang them up (the basketball court) is kind of out in public. I don’t really mind people watching me workout, but occasionally it’s a little awkward.

Me at the top of an inverted row using gymnastic rings
I neglected to get a photo today, but this is pretty much what I looked like.

Especially during the winter, I’ve largely switched to steel club exercises. I can do them in my study, and they’re a similarly good workout. Besides that, they add an element of rotation which I find useful to support my fencing.

One downside of club (and kettlebell) exercises is that they’re really not practical to do to failure.

Exercise machines are great to do to failure. With little or no risk of injury, you can know that you’ve gotten the maximum stimulus to strength and hypertrophy. With kettlebells or clubs, training to failure means that a big steel weight goes flying out of your hands and into whatever happens to be in the vicinity—a terrible idea.

The rings are closer to weight machines. I use them almost exclusively for three different exercises: inverted rows, dips, and pull ups. Do any of them to failure and all that happens is you don’t finish a rep. I suppose actual gymnasts do various sorts of inversions and such, where it might be dangerous to fail a rep. That’s not an issue for me.

Anyway, this morning I got my rings out, put them up at the basketball court, and did a version of the same circuit I did many times over the past five years. I did three rounds of:

Jump rope

This was mainly just to get/keep me warmed up for the rest of the round. I did sets of 100 jumps, and was pleased to find that I could still do them with just one or two misses in a set.

Inverted rows

These are a way to work the lats. I’ll hope I can do pull ups again soon, but even then I’ll probably stick to inverted rows for half my lat exercises. I did sets of 8 rows to failure, which is exactly what I was going for.

Slamball hunter squats

This exercise comes from Mark Wildman’s slamball program. You powerclean a slamball, do a double-pivot to turn 90 degrees to one side, which leaves you in a stagger stance. Then you drop down to a hunter squat (about mid-way between a squat and a lunge), holding the slamball in front of you. Then you stand back up, double-pivot again to face forward, and drop the slamball. Then you repeat, pivoting the opposite direction.

I was doing it with a 15 lb slamball, which isn’t terribly heavy, but I was too lazy to carry my heavier slamball out to the basketball court.

I did sets of 6 to 8 reps (3 or 4 each direction), which isn’t enough to reach failure on my legs. I either need to do more reps or else bite the bullet and bring out the 30 lb slamball.

Dips

Somewhat to my surprise, I actually managed to do 1 dip! That was all I managed though, so I satisfied myself with doing negative dips for the rest of my reps. But if I can do 1 dip, being able to do sets of 3 or even 4 is not far off.

Hollowbody hold

This was easier than I’d expected, I guess due to the Mark Wildman ab exercise program I’d been doing for a few months now. I did a 45-second hold each round.

I did a total of three rounds, which is what I did for most of my rings workouts in years past. It felt like a legit good workout.

I’m thinking two of those workouts per week, plus two club workouts per week, plus two runs per week (a long run and a fast run), would add up to a great, well-rounded exercise routine. Of course, I’ll also want to do a fencing class as well, adding up to seven workouts per week, which is perhaps a bit more than would be wise, but I don’t see any alternative.

Posted on
by

I’ve been hearing for years about how much trouble Social Security is in, and how pretty soon there won’t be enough money left in the trust fund to pay everyone’s pensions in full, and how we’ll have to raise taxes or cut benefits. That’s almost entirely false.

The Greenspan commission that restructured Social Security back in 1983 got almost everything right (which is why we haven’t needed to change Social Security tax rates, diddle around with the cost of living adjustments, nor change the age at which people retire for forty years). The one thing they got wrong?

Back then, about 90 percent of all wages were subject to Social Security payroll taxes. Today, that’s dropped to around 82.6 percent as more income has shifted above the taxable maximum.

Source: Actually, Social Security Nailed It In 1983

The most common suggestion for “fixing” Social Security is to get rid of the ceiling on the amount of income subject to the tax, but that’s the wrong way to think about it. Getting rid of the ceiling would decouple the size of the eventual pension from the size of the income that earned it, which would give conservatives yet another hook for criticizing the program.

The right fix is to boost incomes of those at the bottom, so that once again 90% of all wages are under the Social Security tax ceiling.

Making sure that lower-income people earn enough money to live on will fix Social Security as a side-effect.

Pretty cool, eh?

Posted on
by

In the run-up to the campaign I saw reports of young people, frustrated that Biden hadn’t managed to do the huge student loan forgiveness that he’d tried to do, say that they weren’t going to vote for him or for Kamala. “If he can’t get this thing done, why should I support him?” Here’s why:

“Beginning May 5, the department will begin involuntary collection through the Treasury Department’s offset program, which withholds payments from the government — including tax refunds, federal salaries, and other benefits — from people with past-due debts to the government. After a 30-day notice, the department will also begin garnishing wages for borrowers in default.”

Source: https://finance.yahoo.com/news/student-loans-default-referred-debt-200132438.html