Zach Frey

Eclectic Amateur;

Polymath Professional

[Zach Frey @znfrey]

"I have given up newspapers in exchange for Tacitus & Thucydides, for Newton & Euclid; & I find myself much the happier."

-- Thomas Jefferson to John Adams, January 12, 1812

[Zach Frey @znfrey]

@AJA_Cortes All of my adult children have returned home after their college graduations.

This wasn’t failure to launch; this was us providing them a launchpad for their next stages.

The toxic idea that it’s good to make your kids bootstrap from zero needs to die.

[Zach Frey @znfrey]

Long may she sail!

[Guns&Gadgets @Guns_Gadgets]

The USS Constitution, aka Old Ironsides, is the world’s oldest commissioned war ship still afloat!

She was launched in 1797, one of six original frigates authorized for construction by the Naval Act of 1794 and the third constructed. The name “Constitution” was among ten names submitted to President George Washington by Secretary of War Timothy Pickering in March or May for the frigates that were to be constructed.

Happy Independence Day America! 🇺🇸🫡
#america250 #america #independenceday #july4th

[Guns&Gadgets @Guns_Gadgets]

The USS Constitution, aka Old Ironsides, is the world’s oldest commissioned war ship still afloat!

She was launched in 1797, one of six original frigates authorized for construction by the Naval Act of 1794 and the third constructed. The name “Constitution” was among ten names submitted to President George Washington by Secretary of War Timothy Pickering in March or May for the frigates that were to be constructed.

Happy Independence Day America! 🇺🇸🫡
#america250 #america #independenceday #july4th

[Zach Frey @znfrey]

@Mjohnstone68 @bscholl That’s a slow-motion *tragedy*.

[Edward Mehr @EdwardMehr]

Also why in high mix manufacturing focus on cycle time for a process is stupid. You will always be as slow as your slowest operation which in high mix is the design. To achieve high takt time in high mix environment you have to scale horizontally (speed up with replication)

[Zach Frey @znfrey]

Happy to have Uncle Bob confirm this.

I’m not quite old enough to have experienced the “C compilers produce ASM slop!” argument myself, but I’ve heard. And I am old enough now to have gone through some iterations of the argument myself.

I expect this is what will happen every time a new level of abstraction becomes available.

[Uncle Bob Martin @unclebobmartin]

For the last fifty years it’s been about preventing the slop from overflowing. It was about then that we stopped looking at the true state of the system. We simply assumed that whatever the state of the system was, at the binary level, it was probably good enough.

Believe me, the debate we are currently having over AI slop is the same debate that we had 50 years ago over the difference between assembly language and C.

[AJAC @AJA_Cortes]

America

[BowTiedYukon @BowTiedYukon]

"It is terrible to contemplate how few politicians are hanged."

G.K Chesterton

[Zach Frey @znfrey]

One of these days, I must learn to read the omens

[Hernan Cortes @CyberPunkCortes]

Eagle and Fox together in Dutch Harbor. They both agree something needs to be done.

[Zach Frey @znfrey]

@SteveSkojec

[Zach Frey @znfrey]

I’m not at Carmack’s level, more of a clever ordinary mortal with a few years of perspective.

I have no idea whether prompt-to-binary is a good idea or not. I’m sure there will be any number of PhD students giving it a try to find out. (Publish or perish!)

My knowledge of neural nets pre-dates the transformer architecture that is the basis for modern LLMs. However, I get the basic idea that it’s a long probabilistic stream of tokens with a context window.

There does seem to be a mismatch with the abstract syntax tree nature of computer languages. On the other hand, it’s “merely” an engineering problem to transform one into the other.

After all, the Turing Machine’s memory model is one long stream of tokens.

What I suspect will happen is based on what we’ve done before. Assembly code didn’t disappear with C and other high level languages - it just became a lower level abstraction used by the higher level representation.

Same with model-based development. The C code is the intermediate output from a Simulink model - the code generator doesn’t go straight to binary.
Which is *good* - it lets the C/Assembler combo handle what it’s good at, and simplifies the model-based generation.

So what I expect to happen is for us to end up with LLMs tuned to produce “textual problem statement to code” transformations, outputting Python or Rust or Typescript or whatever, and those languages becoming to mainstream software development what Assembly code is today. A specialized tool that only a few übergeeks use when it’s necessary to look “under the hood.”

It’s certainly “interesting times” in the industry. Time will tell!

[Devon Eriksen @Devon_Eriksen_]

Okay, looks like we're going to need some Carmack-to-ordinary-mortal translation. So I'm going to take a few pages to explain what he said in a paragraph.

When LLMs read code, they parse it as a linear stream of language nuggets. But computer languages are basically designed to describe a sort of tree-like structure of nested meanings called an "abstract syntax tree".

An LLM can emulate an understanding of this by its trained ability to parse specific characteristics of the language, but that only goes so far due to the limitations of the model.

Humans, by contrast, have the ability to actually construct pieces of the tree in our heads, giving us a much more exact understanding of code. But we're limited, because we're slower, especially in typing, and because we can remember fewer things at once.

Carmack suggests that it might be worthwhile to train an AI type that reads abstract syntax trees directly.

Such an AI would first invoke the front end of a language compiler, to parse code into an abstract syntax tree, just as in the process of compiling code to binary. It would then operate on the AST itself, rather than on human-readable code, and then use a reverse-compiler like process to turn its result back into code, for human reading and review.

How transformer architectures might work on a tree is a little bit of a head-scratcher. I'd want to play with graph transformer type models, but I'm reluctant to say anything more than that off the cuff, because my knee-jerk approaches to the problem might turn out to be egregiously stupid.

Would this be useful?

I don't know, but Carmack thinks it might be worth a try. However, it might turn out that linear stream transformers have an emergent "understanding" of tree structure if you throw enough GPU cycles at the problem.

A highly correlated question is how human brains code. I haven't looked deeply into this, but Grok claims that fMRI studies show that human coders do not primarily rely on language centers (Broca's area, etc) for coding, but show a lot of deep-brain activation.

If it isn't hallucinating, Carmack may well be right.

[Fenix Ammo CEO @FenixAmmoHR]

This is because you live in a suburban hellhole surrounded by perfectly manicured, pesticide filled golf course lawns.

[Echoes of War @EchoesofWarYT]

The most influential immigrant group in American history is the one nobody argues about, because almost nobody remembers it was them.

Start at the beginning. The Continental Army was a half-trained mess until Baron von Steuben, a Prussian officer, showed up at Valley Forge and drilled it into a real fighting force. The freedom of the press you take for granted traces back to John Peter Zenger, a German immigrant printer whose 1735 trial established that you can't be jailed for printing the truth. German-Americans were shaping this country before there was a country.

Then look around your own life. Your Christmas tree is German. The hot dog (Frankfurt), the hamburger (Hamburg), the pretzel, the delicatessen, all German. Kindergarten is German, the word and the idea, brought over and opened by Margarethe Schurz. Blue jeans came from Levi Strauss of Bavaria. Heinz ketchup, Steinway pianos, Oscar Mayer, and the big four beers, Budweiser, Pabst, Miller and Schlitz, were every one founded by German immigrants.

The Brooklyn Bridge was engineered by John Roebling, born in Prussia. The Santa Claus you picture every December, plus the Republican elephant, were drawn by Thomas Nast, a German immigrant. Pfizer was founded by Charles Pfizer, who arrived from Germany in 1848. Boeing was built by the son of a German immigrant. John Jacob Astor showed up from Germany with next to nothing and became America's first multimillionaire. Charles Steinmetz, a disabled immigrant nearly turned away at the border, went on to make modern electrical power possible.

And it kept going. Wernher von Braun designed the rocket that put America on the moon. Einstein was German. Carl Schurz, a refugee, became a Union general and the first German-born US Senator. Eisenhower commanded D-Day and won the White House under a name once spelled Eisenhauer. Babe Ruth was a German-American kid from Baltimore.

Here is the kicker. German is the single largest ancestry group in the entire United States, around 44 million people, bigger than Irish, English or Italian. The biggest thread in the whole American fabric, and somehow the quietest.

They never asked for parades. They just trained the army, freed the press, engineered the bridges, founded the companies, built the rockets and lit up the Christmas mornings, then blended in so completely you forgot they were ever the "other." That might be the most American story there is.

[Zach Frey @znfrey]

@ZubyMusic Remember to: breathe! And say to yourself “this too shall pass”

Which is both comforting and tragic.

Enjoy the time!

[Zach Frey @znfrey]

@GPrime85 Well, crap.

I really wanted to like the new Supergirl movie, but if they’re resorting to “the fans are just MISOGYNISTS!!!” already, it must be bad.

[ƨꭋɘƨꭋɘvdo @obversers]

I realized that something very rotten was brewing in a lot of people's outlooks, when I was in university for urban studies and I'd come back from visiting family in Ohio. some people would react like I'd taken a road trip to the Fourth Circle of Hell. meanwhile, a lot of them...

[Zach Frey @znfrey]

@MatthewChang Carbon-neutral too!

[Zach Frey @znfrey]

@zanehengsperger @KellyGibb2506 most affected

[Zach Frey @znfrey]

Best apologia for Imperial measurements I’ve seen. Read and enjoy.

[Devon Eriksen @Devon_Eriksen_]

Okay, time to explain the Imperial system, the metric system, and why attempts to replace either with the other are all retarded.

They have two different purposes.

The metric system is designed around precise measurement of objects. Its goal is to make engineering and scientific calculations simple.

The Imperial system is designed around humans. Its goal is to make calculation unnecessary.

100 degrees is really hot. 0 degrees is really cold. Anything that starts with a 5 is cool, anything that starts with an 8 is warm. No computation.

6 feet is tall, 5 feet is short.

100 pounds is light, 200 pounds is substantial, 300 pounds is heavy.

A 1000 square foot house is small, a 2000 square foot house is medium, a 3000 square foot house is large.

1 mile is a short walk, 2 miles is a medium walk, after that it takes a while.

1 acre of land is a homestead, 10 acres is an estate, 100 acres and up is a ranch or a farm.

Do you see now why it is so strange and awkward to convert from miles to feet?

It's because converting from miles to feet is not something you're supposed to do in the first place. Yes, they are both measures of length, so they are technically convertible, and yes, on rare occasions, you might need to do that.

But feet are for measuring humans, and things built around humans, like doorways, and mattresses. Miles are for measuring travel distance.

You wouldn't measure the distance between Seattle and Portland in feet for the same reason you wouldn't measure the distance between Tokyo and Osaka in mattress-lengths.

It would be silly.

This is why Americans so fiercely resistant to any notion of "conversion" to the metric system. Because it makes no sense. We already use the metric system for what it's good for, which is doing physics and chemistry and whatnot.

But converting everyday measurements to the metric system would be less useful, generally inconvenient, and serve no purpose other than to make petty government bureaucrats happy that everything is now tidy, orderly, and worse, three qualities that bureaucrats love.

I thought about this carefully when I wrote my first science fiction novel. In the world of the 22nd century, extraterrestrial settlers ("Orbitals") use three systems of measurement.

They measure themselves in feet, inches, and pounds.

They measure the spacecraft and habitats they build in meters and centimeters, grams and kilograms.

And they measure space travel distances in light-seconds and light-minutes.

Each system has its own natural scale.

The sole exception to this is when Marcus doses himself with drugs for high-g resistance, Miranda objects that he has taken too much, and Marcus responds by stating his mass... in kilograms.

Why?

Because they're talking about drug doses, a engineering measurement. Drugs are dosed in milligrams per kilogram.

So, yes, the Imperial system makes perfect sense when you understand what it's for, and no, we ain't changing.

And, as a general rule, when an entire civilization of smart people does something for centuries, and it makes no sense to you, they're probably not being silly.

It's more likely there's something you don't know.

[Zach Frey @znfrey]

@EduardHabsburg @SketchyMouse It will be a sign