coding along LLMs, July

AI technology

July 2024 and it is still surprising how erratic LLMs can be when they get tasked to help with very small coding jobs. I work on many projects on my Mac. For a while, I have been using my own directory simple stack implementation. Remembering paths is what the computer can do for me.

It works well and has two parts: zsh functions loaded via .zshrc and a Python program doing the actual work, naturally except for the cd, change of the prompt.

I thought it would be nice if my ‘cd’ variation could check if a Python environment under */bin/activate would exist in the directory I changed into. If so, it can source it. If there is none, then it should not care, and if there are multiple, it should list them so that I could pick and choose.

Simple enough.

Parts would require zsh shell coding. Not something I tend to do a lot. Since Sonnet 3.5 has a limit even in the paid version, I tend to use my paid gpt4o first.

For this simple thing, I should not have. Today gpt4o was stunningly stupid. It managed to do zsh syntax well enough, but then completely failed. For a while, stuck in that dreadful loop where one hopes the next version would finally work. I still abort those loops of idiocy way too late.

Claude 3.5 got it right. In my frustration, I also had introduced a bug / typo on my end. Both gpt4o and Claude would have pointed it out easily if they had seen that part of the code. Claude stood out since its debug hints let me see what I had done wrong. That was beyond my current expectation horizon.

Speaking of: I am amazed how dumb LLMs still can be. Today gpt4o was utterly stupid. Not sure why. Is it zsh that it is not familiar with? Did the system prompt that got assigned to me, or my region, suddenly change? Who knows.

It must be hard to make a living based on some expectations from LLMs. They are really awesome, but can fall off a cliff at any point. Pretty much the opposite from computer work in general.

I expect that people develop all sorts of Cargo Cults in their work with these tools.

Meta Status

AI economy history internet technology

What does Meta do? It turns people into money. Those that are on the Internet, that is—not in a Soylent Green kind of way.

At least, that was the mantra up until 2018. Then Cambridge Analytica broke. And the Q2 2018 earnings gave an inkling of the possibility that not a fixed—and also rather large—ratio of people entering the Internet would become, just like magic, Facebook users.

Later, people seemed to forget about the fact that they get algorithmically nudged in Zuck’s wonderland every step of the way. Wall Street itself realized that revenues at 1 Hacker Way actually kept on rising—until they jumped in 2021. COVID, remember?

The Metaverse, however, wasn’t really that great of a hit, and after the virus bonus revenue fell back in line the following year, FB lost a staggering two-thirds of its value. A trillion-dollar meme stock.

An attribute that it then turned into current heights via hitching itself to the AI bandwagon.

Releasing the LLaMA weights is undoubtedly a commendable move. It sounds utterly impressive when you can claim, “While we’re working on today’s products and models, we’re also working on the research we need to advance for LLaMA 5, 6, and 7 in the coming years and beyond to develop full general intelligence,” in an earnings call. Pretty much like that strange man proclaimed five years ago: “I want 5G, and even 6G, technology in the United States as soon as possible.” Numbers: They go up, up, and up.

Hype aside, I am not really aware of any practical applications for LLaMA 3. Zuck bought lots of GPUs. Both Jensen and I are happy about that. Maybe they thought they had all this data that people have entered in their apps. Maybe they could train a LLM on it. With GPT-3, there was this notion that the size of the training corpus was all that mattered. After all, OpenAI’s chatbot was such a wonder, and it jumped into existence just via the increase of its training data. I speculate that a trillion training tokens derived from FB discussions yield surprisingly little meaningful reasoning power. Especially compared to actual content like, for instance, Wikipedia.

The pressure to come up with something must have weighed heavily on 1 Hacker Way. As those two transformer-based applications (LLMs and Image Diffusers) broke into public view and kicked the world into a frenzy that seemingly became the new normal, Meta itself had just spent around $50 billion on developing, well, the Metaverse. Which received rather little positive reaction, to put it mildly.

The total and utter failure of Zuck’s idea to come up with a whole new thing left Meta with no choice but to jump on the AI hype PR scheme. And up to this day, it has worked rather well. While revenue is ticking along as expected, the stock is kissing new heights. For now.

So, what’s next?
Nobody knows.

What will happen is that Internet population growth will end. There are simply no more people left that could join. Pretty much everybody who could go online already has done so. While 25% of the world’s population are younger than 15, many of them live in underdeveloped parts of Africa. Furthermore, young people hardly flock eagerly into the Meta family of products once they get their first Internet device.

Meta’s revenue growth would therefore stall together with the plateau in its user count. While they continue to make a lot of money, a PE ratio of currently around 30 is expecting something else: More money. You need to grow profits to justify such a valuation.
A quick way to bump revenues would be to reduce costs. Twitter is still up and running, despite Mr. Musk letting go of most of its workforce. A tempting move that could save the numbers for a quarter or two at Menlo Park as well. The problem is that this approach works only briefly: Costs go down to zero. But not more.

Which means that Meta needs to increase revenues while user numbers can no longer grow.

Can Zuck’s companies accomplish that? They might, but it would not be pretty: Billions of people have delegated a great part of their social existence into the “Meta Family of Products”. (What’s in a name?) A sticky situation in itself. Add to that the addictive aspects that rival nicotine, and you realize that half the planet as a user base won’t go anywhere fast.

Wealth as well as the inflexibility to change app use or social topology both tend to grow with increasing age. Meta owns people’s time and attention in staggering amounts.

Here comes the part that isn’t pretty: it is rather easy to manipulate people online. Tech is able to do it. And will increasingly be. There is a threshold after which you no longer realize that you got nudged.

When the magician manages to direct your attention successfully, all sorts of things are possible. With a serious difference: Magic lives from the effect, that the outcome shows you, that you must have missed something. You are supposed to notice that it is impossible what just happened.

Manipulation to gain, aka advertisement, has a different aim: You should be made to act in certain ways, all the while thinking that you want to do that.

The total spending of Meta family users is responsible for a mind-boggling share of GDP. And, as discussed, most of the users will not go anywhere. If Meta does not f*ck up royally, pretty much half of the global adults will continue to point their noses, eyes, minds, and wallets its way.

Turning on the manipulation engine will not be one deliberate conscious act or one magnificent large piece of software. Lots of little changes will yield lots of little benefits. With billions of people, you can do a whole lot of A/B testing. Nobody will notice. Everybody’s feed is different and the fact that you see wording that is ever so slightly different will not trigger any of the societal mechanisms that will raise a reaction.

Jacob Riis used flash photography at the end of the 19th century to show the world how poor people lived in NYC, and he changed the world for the better. I cannot imagine how we can illuminate the modern plight of getting nudged into an ultimately unhappy existence that looms on the horizon.

Power Shower

daily life economy marketing

Orwellian, isn’t that what you call it?

What happened? Some bean counters at Best Western figured out how much money gets spent on heating water for the guest showers. Probably a fair bit, since BW Hotels operates 320,000 rooms globally.

So let’s put in those water-saving devices.

Yes, actually taking a shower is pretty much impossible now. Your body still gets wet. Somewhat, in some places. But forget about mundane actions like washing the soap away from the skin. That measly drip coming out of the device will not do that. Water is an essential feature of taking a shower.

BW Hotels think that this is not a given.

The amazing thing is, that somebody thought to put up a sign reading:

Dear Guest,
Please be
advised that
this shower is
with a gentle,
style head.
It is not a
“Power” shower

Now we know. It is not a bug, it is a feature. Bugs are unintentional. This one very much is intentional. And it sucks. Really badly.

So, if I have the choice to stay in a Best Western or in any other establishment I know what to do.

Trying to reframe what is normal into something special, that got replaced by something special, is evil. Intentional, and I will not support it.

The TV in the room was from 2011. A lot has happened in TV tech since then.

Navigating LLMs: Benefits and Drawbacks in June 2024

AI google history internet

LLMs have their limits, and where they excel makes a difference. As of June 2024, they continue to evolve. Anthrop\c Claude 3.5 works well for coding simple things with Python. It feels like the LLM has been heavily trained on existing code. Actually, it might be just as good in other applications as well. I wouldn’t know since I only use it for coding right now. Even on the paid plan, it has a message limit, which feels very 2023. So, I use the limited interaction where I get the highest benefit, which is coding. The artifact window is a great idea, and the speed of generation is appreciated. With gpt4o, I had to interleave work: make a request, switch to a different task while gpt4o sputtered out characters at Morse code speed. It probably runs on a colony of squids at the bottom of the Mariana Trench that OpenAI taught how to use Morse code with each arm.

And yes, an image like this I create with gpt4o. I don’t even know if Claude can do that. I don’t mind having multiple LLMs. I am gladly paying for both of them, as I do for search. Right now, I am very happy that there is more than one solution. I tried to use Google AI, but it was too complicated to figure out. To find the offering that fit mit my needs. And I am not aware of a key feature that I could only do with them. They already have all my email, read the entire Internet. If I can avoid it, I would not like to help them any further. Sure, if they were as good at coding as Claude, I would use them in a second. I have morals, but I cannot save the world single-handedly either.

One of the bigger fears I have is that LLMs might take the same turn that Google Search did. It was a great idea. It worked great, allowing for a phase of the Internet in the early 2000s that was very promising. Then it became what we suffer from today—a swamp. Barely functional. Generating around $150 profit for Google per user annually. Which means companies make more. Which means that I loose even more than that. The costs of using Google Search by being manipulated are much higher now than its benefits. The SEO world that Google Search presents is not a nice one. I happily give Kagi money to have some distance from that swamp.

Goethe: Kunstwerk des Lebens

books history

Rüdiger Safranski schreibt über das Leben und damit ja auch Werk Goethes. Frank Arnold las “Schuld und Sühne” ganz hervorragend. Das gab den Ausschlag für dieses 28 Stunden Audiobuch. Mir gefiel es gut. Goethe lebte von 1749 bis 1832. Zeiten, in denen viel passierte. Auf Buchseiten und im echten Leben. In dem von Goethe und um ihn herum.

Der Autor bringt dieses sehr komplexe Geschehen sehr wohltuend in eine Form, die es aufnehmbar macht. Ich bin mir sicher, dass am Ende alles in Wirklichkeit noch viel verwobener war. Aber die Lesbarmachung dieser Zeit und dieses Lebens ist – in meiner Sicht – komplett gelungen.

Schnitt, van de Laar

books history

Arnold van de Laar schreibt in “Schnitt!” über 28 Operationen der letzten Jahrhunderte und in einigen Fällen auch Jahrtausende. Das Buch ist nicht unbedingt etwas für schwache Nerven. Als Chirurg vergisst Van de Laar mitunter, dass nicht alle Menschen seine nüchterne und distanzierte Sichtweise in Bezug auf Körperversehrungen aller Art teilen können. Dem Inhalt tut das aber keinen Abbruch, wahrscheinlich sogar im Gegenteil.

Die Unterteilung in verschiedene Operationen, Operateure, Patienten, Zeiten und Regionen, die hier zusammen beleuchtet werden, tun dem Werk sehr gut. Van de Laar beleuchtet die Geschichte durch seine eigene, dabei zugleich immens einsichtsreiche Perspektive.

Seitdem es Zeitungen gibt, füllen sie sich zu 90 % mit schlechten Nachrichten. Ein Buch wie Schnitt! macht in klarstem Maße deutlich, dass wir durchaus schon weit gekommen sind. Wenn einem etwas fehlen sollte und man sich die Epoche der Behandlung aussuchen könnte, wäre die Auswahl immer sehr leicht: Heute. Ganz unbedingt heute. Vielleicht morgen. Aber auf keinen Fall in der Vergangenheit.


AI Command Line history unix

Since its inception in 1977 awk enjoyed a user base of thousands of people.

Since LLMs can tell us how to use it now, it suddenly became millions.

Eine Art U Zentrum für Anspruchsvolle

communication history media

Was auch immer das Internet mit der Menschheit macht, es ist auch eine Plattform, auf der viele historische Kulturformen weiter existieren können.

Radio-Hörspiele zum Beispiel. Als der Massenkonsum schon längst zum Kino und dann zum Fernsehen umgezogen war, entwickelte das Radio seine vielleicht spannendsten Inhalte. Wie einige Podcaster gerade wieder neu entdecken, erlaubt eine Audioproduktion mit relativ einfachen Mitteln die Erzeugung von interessanten Szenarien.

1969 ist lange her. Hermann Ebelings Hörspiel Der Konzern hat auch 2024 inhaltlich noch Relevanz.


data cönversion


He showed the kind of reliable presence that you’d expect from an eldest son of an eldest son.

I really enjoy looking at random blogs. Instead of having something algorithmically pushed in my face, I just stumble upon whatever it is. And deal with whatever it does for me and with me. That this happens outside some intent is a huge plus for me. Like the sentence above. Finding it triggered a question about my ancestry.

In 2008, I had entered data into PhpGedView. The problem is that this software didn’t really get any significant updates. PHP8 frowns upon array{} for instance. The database dump could be loaded, and in the end, it was surprisingly little trouble to extract the GEDCOM records out of the different tables into a GED file. Webtrees looks like a reasonable application. The installation was nicely guided. After a quick expansion, the system nicely indicates which modules are missing, which permissions should be different, etc. All very easy to solve. 25 years ago this kind of thing might have taken half a day with varying success. Even though technically software was already flowing via the Internet back then. The import of the GEDCOM data itself worked fine.

Traveling as an analogy is helpful to indicate how insane this all is: If things go well, like they did here, one travels at jet speed. Around 1 km per breath taken. Nice. Great that the data is now in a neutral GEDCOM format, will make it into a backup, and exists in living software.

Just some umlaut flaws to fix. Maybe 20 entries that don’t look right. I could have fixed them manually. But NO, why not write a quick converter instead? After all, GPT-4 was so helpful before. This should take 5 minutes.

If man makes a plan, God starts laughing.

It took 4 hours. I ended up willing a stupid search-replace tool into existence. Because, well, GPT-4 completely lost the plot on this very simple problem. To an extent that it seemed to make progress, but then just entirely failed again. It was ridiculous, infuriating, beyond belief.

Or to stick to the travel analogy: It felt like being duct-taped to a garden chair in a Walmart parking lot at midday in August in Arizona.

The first part was so complicated and worked so well. The second one, simply going from èu to ü, completely failed. In a bad way. I ended up doing it myself. Of course, writing a program instead of quickly editing those 20 entries.

This on the background that on May 24, 2024, GitHub CEO Thomas Dohmke said that there will eventually be one billion developers on his platform. Developers, Developers, Developers.

random weblog

internet weblog

I didn’t find that mythical UrBlog of mine that I mentioned in the first post of the recently recovered one. The loss of something digital is weird. Since it lacks the inevitability of entropy that all other things have, it is especially infuriating. It could have been prevented.

Oh well.

During the search, I came across some other things from the past. I had written a ‘meme tracker’ in the early 2000s called BlogsNow. It would reliably detect which links gained popularity. Back then, the so-called “Blogosphere” was a crazy mix of all sorts of things.

Finding the data for it got me thinking about the status of those 7 million weblogs. How many would still be updated after two decades? I had just seen myself that it is not easy.

There are around 3500 site left that still get updated regularly. I didn’t spend too much time on the parser. There is a margin of error. But one in two thousand is a pretty strong filter.

When clicking around, I was surprised that some of the pages are rather interesting and surprising. So I added a random blog link on my main page.