atombender

ParadeDB is not "another DB", it's a Postgres extension.

Sorry to break it to you, but there are many, many novelists who prefer to write longhand!

Some notable examples I remember include J. R. R. Tolkien, Patrick O'Brian (who wrote all 21 books in the Aubrey-Maturin series by hand, and claimed at one point to write with "ink and quill"), Nabokov, Joyce Carol Oates, Ursula LeGuin, and Truman Capote.

People work in different ways, of course. From my perspective, agents have gotten so good that there is no reason whatsoever to look at what they are doing while they're doing it.

I think we are moving in the direction where agents work a lot more autonomously. Most of my work these days is done by cooperative teams of remote, autonomous agents. They can work for hours and periodically reach out to humans with questions and updates. But there's rarely any point in looking at their work (mostly diagnostic or auditing purposes). So for me, at least, the days of staring at the agent's output are long gone.

But I still use local harnesses for some tasks, and I'm not happy with OpenCode's desktop app. I'd be nice to have a viable alternative.

Examples?

Hi, I tried Juggler a few months back, and while it's technically impressive as a one-person product, I found it super stressful to use. It emits so much output and activity, in various nested forms.

Where Claude Code and Codex have made progress recently, I feel, is with sharpening the interactivity. For example, Codex will emit a lightweight ream of "thinking", then on completion it will erase this and only show the final, coherent answer. It doesn't show much detail about which commands it runs, either. You can still drill into that, but it's usually the least important thing. Claude also does sub/side-sessions that can talk to the other sessions. So if a session finds an unrelated, pre-existing bug, it can start an adjacent session for the bug fix, and keep abreast on when the fix lands. If you start a side session to do research, it can message the other session with new findings.

I don't know if Juggler is still like it was, but I strongly recommend making it less verbose. I just couldn't deal with the noise.

This is unnecessary whataboutism. The assertion I was addressing was "student-led revolutions have never produced a long-term positive outcome", not "all student-led revolutions have produced a long-term positive outcome".

There are several examples of positive student-led revolutions. In the Velvet Revolution in Czechoslovakia, the student-led protest organizations were the driving force for change. [1]

The mass protests that toppled the military regime in South Korea in 1987 was also led in large part by students and faculty members, who later became prominent leaders. [2]

This has also happened twice in Serbia now. Milošević was defeated in 2000 thanks in large part to Otpor, a student-led organization [2].

[1] http://news.bbc.co.uk/2/hi/8363827.stm

[2] https://en.wikipedia.org/wiki/June_Democratic_Struggle#Stude...

[3] https://www.wilsoncenter.org/event/bringing-down-dictator#1

It's not taking a photo. It's computation, based on a lot of real data, with a lot of inference and synthetic data. The idea is that it's broadly accurate but cannot capture fine detail.

The approach has been calibrated against known objects like NRAO 530, a quasar near Sgr A* that has some known properties, and it's been used on both Sgr A* and M87. And the team has admitted there's a lot of uncertainty about exactly how accurate it is.

Here [1] is a pretty good article about how it works.

There's overwhelming scientific consensus that Sgr A* and M87* are black holes. It's not really provable at this point, but I don't think anyone doubts it.

[1] https://www.discovermagazine.com/the-event-horizon-telescope...

It's human-facing.

Humans cannot be in the loop for autonomous work. I have agents review code and design/architecture documents, write alert messages based on rules, write analyses (e.g. outage incidents), and many other things. "Having a person write words" would mean not using agents, defeating the whole point.

I do not use agents to write on behalf of myself. I write my won technical write-ups and my own slide presentations. A lot of people don't; and it shows. There's much that agents can do better than humans now, but not everything.

Do you then recommend Sol for technical documentation?

What's confusing to me in addition to the naming is that there are no published guidelines I can use, as far as I can see. Put yourself in the user's shoes. On what basis should one pick a model for a given task?

I can understand smaller models being dumber and faster. That aligns well with scenarios where you are willing to compromise; e.g. you're classifying content or asking a model to format text into a table.

But when you want "maximum intelligence, cost and time be damned", there's the implication that bigger models are smarter on a linear scale, and every provider's model release naming is currently conflicting with that expectation.

This isn't just a naming problem, but a communication problem. If I knew that Astra was the best at coding but a bit bad at writing, and Sol was almost as smart but better at technical writing, then that's the kind of thing I expect to be documented.

One possible way: A simple matrix with horizontal columns like "writing", "coding", "math", rows are the models, and each box is a number of stars from 1-5.

Instead, this sort of thing becomes lore, shared by word of mouth.

@dang, could we replace the link with this much better article?

https://www.theguardian.com/environment/ng-interactive/2026/...

OED, Merriam-Webster, and Collins all show a short "e", as in "delete" or "different", i.e. /dɪ/. I've never heard it pronounced as /diː/. Do you have some kind of southern drawl where you put the stress on the first syllable, like some people do with "defense"?

California wants a ban because they specifically think safety standards are not sufficient [1]:

  California’s ban would be an acknowledgment that the state’s laws
  requiring safer processing techniques for quartz are not enough to
  protect workers: Compliance is too low and the industry too
  fragmented, with many small shops that escape inspection, according
  to the California Department of Industrial Relations, which has
  pushed for the ban.

[1] https://www.nytimes.com/2026/10/05/us/politics/quartz-counte...

What amazes me is that people who write these articles must spend so little time reading them. Or are people just so blind to the quality of writing that they don't notice?

For the most part I do my own writing, because AI models aren't even close to reaching human quality. But for things like dynamically generated technical documentation or automated tasks (like summarizing the results of an automated anomaly investigation), where I cannot be in the loop, I find writing can be improved by asking for a specific style.

For example, if you ask for Simplified Technical English, we go from:

    Storing a lake as thousands of 1 MB Parquet files is a bad practice anyway, and 2.0 does not rescue it. Fundamentals still matter!

To:

    It is bad practice to store a data lake as thousands of 1 MB Parquet files. Version 2.0 does not correct this. The basic rules still apply.

Claude loves gerund phrases where the important stuff ("...is bad practice") is at the end, rather than the beginning ("it is bad practice to..."). But even if "x is y" is logically simpler, I think "it is y to x" is cognitively more intuitive to humans, especially for longer sentences where the x is a long explanation, because it mentally front-loads the topic under one heading: "bad: <stuff>" as opposed to "<stuff>: bad".

Claude's prose causes fatigue, I think, in part because the grammatical structure is at odds with how human cognition expects to be served information in their native language. But I also suspect that's a training thing. We've not grown up reading or writing like Claude.

Yep, agreed. I don't really understand why 1stdibs even exists, since I can't imagine anyone ever pays these prices. Chairing and Pamono (now owned by Chairish) also is generally inflated.

My go to sources are Selency, Vntg, Instagram, Catawiki, and then a bunch of online stores across Europe I've bookmarked, including auction sites like Bukowskis in Sweden. The smaller vintage stores are the ones to go to generally.

In Aliens, there is already a colony where they land. The colony has roads (the original prologue, which is in the special edition, shows Newt's dad driving a truck) and presumably a landing area.

In Alien, the planet landscape definitely is chaotic and hostile to landing, and indeed the crew's ship is damaged during landing.

No, but the courts seem to agree with them.

I use a frosted LED bulb and don't see any reason it would work better with an incandescent bulb, i.e. it works great.

Alas, a good-condition Tizio usually sells for around $200-$350 used, depending on the exact model. First-generation models may go a little higher, like $500-$600. I bought mine a couple of years ago for about €250.

If you browse sites like 1stdibs or Pamono, you're going to see dramatically inflated prices, but I don't see why anyone would ever buy there. Ebay, Selency, Vntg, etc. are more likely to sell at actual market prices.

Len actually wrote a whole bunch books about humans meeting inscrutable aliens — Fiasco, The Invincible, Eden, His Master's Voice — but I think the best one is still Solaris. Many people know the film adaptations, but the novel is ultimately somewhat different in motivation and conclusion. It's a great read.

(One thing I just realized: For some reason Lem loves starting the book with a landing. All of these except His Master's Voice start with the main character(s) either landing a spaceship or docking with one.)

Blindsight is fantastic. The sequel, Echopraxia, is also excellent; it is very different, and I expect many people were disappointed that it wasn't just more of the same.

Also a big shout out to his less well-known novel, The Freeze Frame Revolution. It's about the crew of a "stargate"-seeding spaceship trying to stop the ship's AI, but being hampered by the fact that the AI is both near-omniscient and that they only work in short shifts before being put back in hibernation. He also wrote a handful of short stories set before and after the novel.

There's something very refreshingly direct, unvarnished, no-nonsense, and clear-eyed (some would say cynical) about Watts' style of writing. His prose tends to be tough (quite like early Gibson's noir style, with the punk attitude intact), but also beautifully cinematic. The only other write that I think comes close, aside from Gibson, is Dan Simmons with his Hyperion Cantos, but he's not nearly as darkly cynical.

Just because they don't make much money doesn't mean they want to just give stuff away.

You're quoting an article about a German museum. This is a different museum in France. You can't assume they have anything in common.

Just speculation on my part, but the museum's motivation seems clear enough: they realized that point cloud documents could be used to make extremely high-fidelity copies and thereby threatened the museum's monopoly on official reproductions and other merch, and indeed create counterfeit versions of Rodin's works. The repeated allegations that the author is a counterfeiter with malicious intent suggest that they care about reproductions very much. To me, at least, that explains every single action on the part of the museum.

What is less clear is why the high court, which is supposed to be neutral interpreters of the law, pre-emptively reclassified point clouds as non-document materials that are excluded from FOI requests, contrary to the lower court's judgement. I can only guess that there's some collusion going on at high levels.

Vonnegut's inspiration was Nobel prize winner Irving Langmuir [1], apparently.

In my opinion, Hoenikker is just too amoral and aloof to really match any one of the Manhattan Project scientists. Teller (who suggested using atomic weapons to excavate harbours) and Fermi (who before the Trinity test took bets on whether the bomb would light the atmosphere on fire) are maybe the closest contenders.

[1] https://gizmodo.com/the-real-life-scientist-who-inspired-kur...

Dr. Strangelove was mostly inspired by Werner von Braun (according to Peter Sellers). Clearly there are bits of many other people in that characterization, not just one.

Believe it or not, some of us actually enjoy reading writing that goes beyond 5th grade English.

I'd like to see AI "democratize" the world, but I have a much more pessimistic outlook. One thing we see with technological revolutions is that (once the dust settles) they provide general improvements for all strata of societies. But as some boundaries are erased, others are magnified a hundredfold wherever robber barons can stake an exclusive claim.

If genuine AGI is developed, its everyday conveniences may continue to be shared with the common masses, but the powers that own the technology and the data centers will have no incentive to give everyone access equal to superintelligence, and that includes giving governments the same access.

Every stage of the industrial age has seen powerful, wealthy people/groups (whether you call them oligarchs or industrialists or whatever) become able to exert disproportionate power over the state, and we're already in a place where this is happening. Humans have shown time and again that once you have some resource (whether it be oil, gold, or data centers), someone will want to own all of it, and that they're always willing fight for that ownership, and this fight rarely benefits ordinary humans like us.

Developed countries have been enjoying an unusually long peace time thanks to the wonders of capitalism and the emergence of stable, progressive capitalist-friendly democracies. The road to utopian basic income and limitless leisure is going to be fraught with riots and violence, and make Marx suddenly utterly relevant once again, because there's too much at stake, and too stark a division between the haves and the have-nots.

The alternative timeline is one where governments get together to regulate AI into some kind of international agreement under the UN. That requires taking power away from the private corporations; historically, that's been a tough sell.