The vast majority of software projects fail. Honestly, I can't remember ever in my career working on a project I really believed in.

Sometimes I do enjoy the challenge of doing the impossible. Turning a doomed project around or at least minimizing damage. I had some where I thought "this worked out but if anyone but me had been in charge, yeah this would have been a disaster". That feeds my ego. Though I never ever get any thanks from management or any praise. Though this is more of a German culture thing.

There is a reason why burn out is so high in software dev. You are set up to constantly fail. If you succeed against all odds you get more and harder work until you fail.

You got to focus on yourself and find joy in the little challenges. Don't fret over things that you can't change.

Vocabulary is just the surface. Beneath it is an understanding of how to achieve your goals with design. How to make things that are easy to use, accessible, that create a certain impression.

Does this website (presumably made with the help of these AI tools) show this kind of understanding of design? Not really. It's chaotic, the text is often hard to read and there is a ton of fluff, both in terms of visuals and copy.

There is a "Frequently Asked Questions" section and a "Popular" $100 tier in the "Support the Project" section, even though this project seems to be brand new. Why lie to the reader?

I still remember the two gentlemen in their black, faux leather jackets who rang our doorbell and demanded to see our dinghy. (dinghies where registered products too) We showed them our dinghy, they said thank you and left.

Probably someone fled over the Baltic sea to Denmark in a dinghy. So the secret police went from door to door until they found someone who could no longer show it to them...

This was in the late 80s.

It causes a mic at the other end of the room to get cut into the DJ's live feed monitor with a semitone shift down and some reverb. This causes all sorts of inner-ear chaos and usually clears a DJ off the stage when they're over time within a few minutes at most -- usually under 30 seconds. One time they were trying to figure out why it wasn't working and discovered that the DJ had muted their monitor feed, which explained why they were not only peaking the meters but over time: They hadn't heard the FOUR warnings from the back of house that it was time to wrap up.

There’s your answer.

6 days means on a long enough enough timeframe the load will end up evenly distributed across a week.

8 days would result in things getting hammered on specific days of the week.

> Palantir is working on a tool for Immigration and Customs Enforcement (ICE) that populates a map with potential deportation targets, brings up a dossier on each person, and provides a “confidence score” on the person’s current address

So essentially, the relevant app here is custom built in order to help ICE raids.

That's substantially different from generic office tech where ICE happen to be one of millions of users.

Probably about the same as the US when it was a developing nation. "How the United States Stopped Being a Pirate Nation and Learned to Love International Copyright":

> From the time of the first federal copyright law in 1790 until enactment of the International Copyright Act in 1891, U.S. copyright law did not apply to works by authors who were not citizens or residents of the United States. U.S. publishers took advantage of this lacuna in the law, and the demand among American readers for books by popular British authors, by reprinting the books of these authors without their authorization and without paying a negotiated royalty to them.

* https://digitalcommons.pace.edu/plr/vol39/iss1/7/

A week ago Scott Hanselman went on the Stack Overflow podcast to talk about AI-assisted coding. I generally respect that guy a lot, so I tuned in and… well it was kind of jarring. The dude kept saying things in this really confident and didactic (teacherly) tone that were months out of date.

In particular I recall him making the “You’re absolutely right!” joke and asserting that LLMs are generally very sycophantic, and I was like “Ah, I guess he’s still on Claude Code and hasn’t tried Codex with GPT 5”. I haven’t heard an LLM say anything like that since October, and in general I find GPT 5.x to actually be a huge breakthrough in terms of asserting itself when I’m wrong and not flattering my every decision. But that news (which would probably be really valuable to many people listening) wasn’t mentioned on the podcast I guess because neither of the guys had tried Codex recently.

And I can’t say I blame them: It’s really tough to keep up with all the changes but also spend enough time in one place to learn anything deeply. But I think a lot of people who are used to “playing the teacher role” may need to eat a slice of humble pie and get used to speaking in uncertain terms until such a time as this all starts to slow down.

I'm not trying to argue Palantir is an ethical company; my views on "company ethics" are nuanced but I wouldn't put them anywhere near my "places I want to work" bucket. But (contrary, perhaps, to their name), they're not some weird deep demonic trove of personal information; that's supplied to them by their customers, which is where change needs to happen.

“The Hungry Tree is an otherwise unremarkable specimen of the London plane, which has become known for having partially consumed a nearby park bench.”

Prolog: "Mistakes were made"

As an avid Prolog fan, I would have to agree with a lot of Mr. Wayne's comments! There are some things about the language that are now part of the ISO standard that are a bit unergonomic.

On the other hand, you don't have to write Prolog like that! The only shame is that there are 10x more examples (at least) of bad Prolog on the internet than good Prolog.

If you want to see some really beautiful stuff, check out Power of Prolog[1] (which Mr. Wayne courteously links to in his article!)

If you are really wondering why Prolog, the thing about it that makes it special among all languages is metainterpretation. No, seriously, would strongly recommend you check it out[2]

This is all that it takes to write a metainterpreter in Prolog:

  mi1(true).
  mi1((A,B)) :-
          mi1(A),
          mi1(B).
  mi1(Goal) :-
          Goal \= true,
          Goal \= (_,_),
          clause(Goal, Body),
          mi1(Body).

[1] https://www.youtube.com/@ThePowerOfProlog

https://www.metalevel.at/prolog

[2] https://www.youtube.com/watch?v=nmBkU-l1zyc

https://www.metalevel.at/acomip/

That's just a different bias purposefully baked into GPT-5's engineered personality on post-training. It always tries to contradict the user, including the cases where it's confidently wrong, and keeps justifying the wrong result in a funny manner if pressed or argued with (as in, it would have never made that obvious mistake if it wasn't bickering with the user). GPT-5.0 in particular was extremely strongly finetuned to do this. And in longer replies or multiturn convos, it falls into a loop on contradictory behavior far too easily. This is no better than sycophancy. LLMs need an order of magnitude better nuance/calibration/training, this requires human involvement and scales poorly.

Fundamental LLM phenomena (ICL, repetition, serial position biases, consequences of RL-based reasoning etc) haven't really changed, and they're worth studying for a layman to get some intuition. However, they vary a lot model to model due to subtle architectural and training differences, and impossible to keep up because there are so many models and so few benchmarks that measure these phenomena.

-- Support for .parquet, .json, .csv (note: Spotify listening history comes in a multiple .json files, something fun to play with).

-- Support for glob reading, like: select * from 'tsa20*.csv' - so you can read hundreds of files (any type of file!) as if they were one file.

-- if the files don't have the same schema, union_by_name is amazing.

-- The .csv parser is amazing. Auto assigns types well.

-- It's small! The Web Assembly version is 2mb! The CLI is 16mb.

-- Because it is small you can add duckdb directly to your product, like Malloy has done: https://www.malloydata.dev/ - I think of Malloy as a technical persons alternative to PowerBI and Tableau, but it uses a semantic model that helps AI write amazing queries on your data. Edit: Malloy makes SQL 10x easier to write because of its semantic nature. Malloy transpiles to SQL, like Typescript transpiles to Javascript.

But I also see the difficulty that Astro faced here. Despite being happy with the framework, I never paid for it. The paid offerings didn’t strike a chord with me. And it was partly because whatever they offered, Cloudflare already offered on a very generous free tier.

I'm glad the team have got a second life within Cloudflare,. I'm happy for the people who've given me such excellent software for free for years. Thanks folks!

I remember that day vividly.

It was the middle of the Great Recession, it was the worst our economy was doing in a long time. Millions were out of work. People were looking to the government to see what the plan was to get the country back on track.

A reporter asks McConnell what the senate’s number one priority was.

The answer? Not fixing the economy, not helping out every day Americans. Not finding the root cause of the crash and making sure it doesn’t happen that way again.

No, the answer was “make sure Obama is a one term president.” That’s all we would expect from the senate for the next 6 years.

The day McConnell said that, I said out loud: “I will never vote Republican again for the rest of my life.” (Prior to that point I mostly voted D but not 100% of the time.)

And I plan to keep that promise until I die.

You can turn the garage into a hangout spot. A neighbor has a full bar with communal table plus TV for sports and he opens up the garage door once a week on a schedule (Sunday game day or whatever depending on the season) and whenever he feels like it on work week evenings. As people pass by we invite them over and after a few months everyone knows that when the garage is open, they can come over for a drink and to shoot the shit. Low pressure social interactions that often turn into weekend outings, regular poker games, etc.

Now years later we get impromptu block parties when he brings out the grill onto the driveway. It’s done wonders for our community in an otherwise unwalkable SoCal suburb.

That said, I’m not surprised to see a list of CVEs impacting devalue. After running into some (seemingly arbitrary) limitations, I skimmed the code and it definitely felt like there was some sketchiness to it, given how it handles user inputs. If I were nefarious or a security researcher it would definitely be a focal point for me.

Maybe you’ll be lucky enough to get picked up so you can get your proof.

Founding documents don't do shit. What one needs is a culture which is perpetually hostile towards power. All problems of power are social problems. No law, founding document, principle is going to prevent people from doing stuff if they want to do it.

By working for them you are enabling this.

The shutdown was initiated by chancellor Gerhard Schröder. After killing Germany’s nuclear sector, he signed off on Nord Stream 1 as he was on his way out of office. Just after leaving office, Gazprom nominated him for the post of the head of the shareholders' committee of Nord Stream AG. Russia later nominated him to be on their largest oil producers board.

This guy basically sold out Germany’s energy independence for Russia.

0: https://en.wikipedia.org/wiki/Gerhard_Schr%C3%B6der