magnio

FWIW, in my limited experience, this mostly holds for undergraduate level, and the gap is much closer for post-graduate textbooks.

Hence why scientific research is not really worse in the US compared to Russia, as dumb as using textbook as a proxy for development can be.

> Who's checking?

Yeah I sympathize with people saying software developers should not be called engineers too.

> On Mac, some command-line tools included with macOS or the Xcode Command Line Tools (including curl, git, python3, ruby, perl, and openssl) check certificates against /etc/ssl/cert.pem. For certificates that chain to a root CA certificate in this file, these tools don't check Apple's block or whether the certificate has been revoked. This file will be updated in a future macOS update. If you use these tools and don't want to wait for the update, you can remove root CA certificates from the file.

What a clusterfuck

Uhm, disallowing third-party stores, sure, but you draw a line that allows a non-jailbroken and updated device to be able to access company data? How is the lack of third-party app stores gonna prevent someone from sending sensitive data through the Mail app or to iCloud? Shouldn't company devices be managed with MDM to begin with?

Another one is Stoat (formerly Revolt): https://stoat.chat/

If it gets better and better for the next 40 or so years, will your career exist?

I'm a bit bumped when I first saw pi is moving from bash to codemode and also adding MCP, since I thought that loses the purity and simplicity of the "use bash for everything" philosophy. However, after reading more about it, I realized codemode is just a slightly enhanced version of bash: more complex, sure, but likely more robust, secure, and efficient. For those who, like me, don't get the point of this change, here is how I understand it.

The first tool execution runtime in harnesses are direct tool calls with JSON or XML, such as the Read and Edit tools. As an escape hatch, we have Bash tool that allows arbitrary code execution on the host running the agent. The downsides of using bash (on the host) as the main tool execution runtime are:

- Syntax and obvious errors only surface at runtime

- Unergonomic orchestration of parallel and background tasks

- Verbose command output cluttering context

- Dependent on the host environment, packages versions, etc.

- No security measures by default.

To me the last point is the biggest inherent weakness, usually mitigated by creating a dedicated unprivileged user or running bash in a sandbox.

Note that direct tool calling is kind of the polar opposite on these points: syntax errors are caught early, orchestration can be done with some wrapping tools, command output is controlled, and most importantly they are more sandboxed. On the flip side, they obviously have way less power, necessitating Bash tool in the first place.

Codemode is the middle ground between these two extremes. It actually can be derived simply by one idea: what if we replace Bash by another language that can be checked for obvious errors, i.e. type checked?

Everything else falls out from there:

- Any language would do, but I think TypeScript fits the balance between safety, speed, conciseness, and popularity in training data.

- If we use TypeScript, might as well run it in a sandbox as JS runtimes have been designed with this in mind for 20 years

- Orchestration comes for free from the JS runtime. It's not more powerful, just more ergonomic.

- Since the tools are controlled by the harness and not dependent on the host, cloud agent becomes easier.

- With this in place, MCP are not very different from a tool provided to this sandboxed runtime.

Overall I find the benefits compelling enough, but we'll see if the heavily-RLed models these days will use it effectively.

The more I read these ass Claude Opus writings the more I believe AGI is further away from us.

I find it's surprising that people doubt the effectiveness of Pangram when I feel like its job has gotten easier ever since Claude models got RLHF to death with its bizarre punchy negatives and tortured analogies, making AI writing much more noticeable than the GPT 4o era.

> He described waking up to thirty PRs every morning, each one pulled overnight by someone's AI, and needing to make snap judgment calls on every single one. The temptation to delegate the review itself to an AI was enormous. But, as he put it: "at that point, what am I still doing here?".

It's so funny and somber to see programmers having an existential crisis when they get a glimpse of what work is like for business managers, the demographics many programmers detest.

I am also guilty of holding the business majors in contempt back in college, and now here I am, doing what they are doing in office in a much more indifferent and unenjoyable manner. At least I don't get into trouble with HR from calling my agents a stupid fuck (yet).

I used Better Auth for my mobile app backend. It works okayish. My biggest complaints are the OpenAPI specification gets little care, as it mainly caters for JS frontend, and breaking changes in patch version are more common than most packages.

I saw on Twitter that in an ML course at Tsinghua University, one of the tests asks students to write quizzes that fail the most LLM models as possible.

What if we create a benchmark that works like this and assigns ELO scores? Models fight head-to-head by writing a question, a bug, or an incomplete implementation, which the opponent has to answer, fix, or finish.

i like the COPY and logical replication improvements. Currently I back up my PG database with a sidecar Databasus instance that is heavier than my entire backend + DB + Caddy!

(LLM writing rant below)

---

> That alone tells you something: Users had a real need, and the ecosystem filled the gap.

> This sounds straightforward, but it solves a real operational problem.

> None of these change the world. All of them make day-to-day data workflows better.

> The easy thing to do here is list planner changes and call it done. But the more useful takeaway is this: Postgres keeps getting better at recognizing the shape of common queries and doing less unnecessary work.

> [Proceed to list planner changes]

If Orwell were alive today, he might declare himself illiterate in English and learn Klingon just to avoid having to read these.

Yeah, the big 3 cloud markup is so high that most VPS providers can hike price 10x and they are still cheaper.

Heck, mise has an HTTP backend that can install binaries from any URL. I use this to manage Atlassian CLI, whose official Windows binary is not on winget.

Yes, and way before vibe-coding is a thing. Back in 2022, a version of node-ipc formatted the disk of users in Russia and Belarus.

https://arstechnica.com/information-technology/2022/03/sabot...

I doubt Go has any sizable effect on the community of programming language developers. Probably Pascal has more impact on this.

Never, ever, ever transform URIs and paths by string manipulation. If you think pulling in a library for this is overkill, it is not.

(Lesson learned from trying to quickly write my own function to make ".." to go back one URL segment that took 3 hours and discovering the URI spec contradicts my intuition depending on whether the URI is a URL or filesystem path.)

To me, the fact that it exhibits various telltale signs of LLMism is not the main problem; it is annoying mostly due to personal preference, just like I am annoyed by the writing style of some authors from the 19th century.

The main problem is that LLM writing inevitably slips in nonsensical phrases and sentences that are plausible but, upon inspection, turn out to be dilutions at best and deceptions at worst. They are such non-sequiturs that it is indefensible to consider them the crystallized results of a logical thought process, so I greatly dislike them regardless of authorship, and so far, it has been mostly LLMs that produce them. However, this is not a new thing, as Orwell put it from 1946:

> the slovenliness of our language makes it easier for us to have foolish thoughts [...] Modern English, especially written English, is full of bad habits which spread by imitation and which can be avoided if one is willing to take the necessary trouble. If one gets rid of these habits one can think more clearly, and to think clearly is a necessary first step toward political regeneration: so that the fight against bad English is not frivolous and is not the exclusive concern of professional writers.

Here are some examples from this article, which I am listing with the sole hope to elicit the same concern of frivolous writing from readers, not as any an attack on the author of this piece:

> The 1955 trade was the system already past its own decision point, picking up the pieces. The blunder happened five years earlier. [...] Everything after, [...], was the system mechanically playing out the consequences of the June 1950 decision.

What is "the system" here? The US government, the FBI, or just a vague blob of fate and the passage of time? What exact had this system been doing that counts as "picking up the pieces"? What was exactly mechanical about the events between 1950 and 1955?

> He later recalled in characteristically dry phrasing that "many of the officers in the U.S. Army in missiles and rockets were students in this program".

Is the quote really especially dry? Is this purported dryness a part of Qian's character? If yes, then we sue did not see it elsewhere.

> The hierarchy in the interrogation room was not what casual U.S.-narrative framings would assume.

What is even "casual US-narrative framings"?

> The Jiang family connection is structurally important, and the 1955 PRC-side claim that Qian was a long-standing Communist sympathiser is structurally implausible because his wife was the daughter of a senior Kuomintang figure

What do "structurally important" and "structurally implausible" mean here? What is the big structure that this connection played an important role in?

> This was the public peak.

The peak of his publicity I guess. The public peak is in Nepal.

> He was, on the public record at that exact moment, one of the leading American aerospace scientists. He was not a junior researcher who could be replaced.

Who read till this point would need this clarification that he was an expendable junior researcher?

> They are produced by a sequence of external shocks that hardens the U.S. political environment around him in the eighteen months before his clearance is revoked.

(Minor but somehow this sentence is in present tense.)

> His later assessment of the trade was accurate pricing made in real time by an official with the position to assess it, though he was reading the wrong decision.

"accurate pricing" as metaphor for "accurate assessment"?

> The imprisonment was the trade's cause, not the trade.

The trade was not the trade's cause, or the imprisonment was not the trade, or something else?

> The Soviets were trying to slow the propagation of capability that had already been absorbed

If the knowledge had already been absorbed, what use is withdrawing the blueprints? The conclusion that the Soviets must have thought the Chinese could become independent because they took the materials away is quite dubious.

> the capability foundation was laid

What does "capability foundation" mean?

> The doctrine is the kill-first-from-distance-using-superior-detection-and-networked-sensors-and-long-range-missiles doctrine that Qian outlined in the Toward New Horizons volume on the launching of a winged missile for supersonic flight.

No mention of the doctrine prior to this sentence, while the next sentence says that the aformentioned PL-15 embodies that doctrine. Could it have just been written as one sentence?

> the structural threat to U.S. naval power projection

"structural" here we go again

> The same pattern is visible at every other layer of the strategic-technology spectrum in 2026.

No mention of any pattern prior to this point, except the vague development of China war capability.

> This is what compounding looks like when you imprison the carrier you needed to retain

Does this event have that many precedents that it deserves to be written down as an aphorism?

> That chain was the thing that walked out the door. The methodology is what the chain was running on

I honestly don't know what chain and methodology here refer to, much less whether the metaphor is sensible.

> The dimension of the transfer that has no Soviet equivalent and no Western parallel is this one.

"The dimension of the transfer"?

> structural features

"structural"

> multi-disciplinary integration across specialties

"multi-disciplinary" = "across specialties"

> The methodology was specifically Western, specifically von Kármán-lineage, and specifically transferable through a single carrier

What does "specifically" even mean when it applies to 3 things at once?

> The fact that he was available to be that carrier was a function of the Boxer Indemnity Scholarship program, von Kármán's recruitment decisions at Caltech, the wartime mobilisation that placed him at the centre of the U.S. air-power apparatus, and the Red Scare architecture that produced his imprisonment. The full chain had to operate. Removing any link in it produces a different outcome.

A lot to say "changing the past affects the future".

> Wang Huning, who became Xi Jinping's chief ideologist, sits in a tradition that runs directly through Qian's cybernetic-systems-engineering work, and the methodology runs into Chinese state-planning architecture along that lineage

What is this "tradition", and what is that "lineage" exactly?

> The way I read it

A subjective thought! A cause for celebration.

> paired with Chen Yun's 摸着石头过河 as the operational sidekick.

"operational sidekick"?

> The same machinery, operating on essentially the same evidentiary basis, produced bounded internal exile for Oppenheimer and unbounded external transfer for Qian

"bounded" and "unbounded" here mean nothing, as both are bounded by the Earth size? What is even the point of talking about geographical difference when the political difference is of dominant interest here?

> structural dynamic

> moral architecture

...

> The Oppenheimer-Qian-Japan triangle reveals the same pattern repeatedly.

Finding these supposedly aforementioned "pattern"s is like finding Waldo.

> structural rights barriers

> US self-conception

> threat-detection regime

> consequence space

> methodological irony

> structural feature

> structural reading

No worries, I don't mean to disparage your article. At least it avoids some of the most annoying LLMism I have seen, and given its length you must have put some effort into prompting, researching, or editing. Hope you will find your own voice as you write more and more.

If you wanna read an article containing essentially the same information without the pesky LLM voice: https://www.usni.org/magazines/naval-history/2025/december/m...

Is diff viewer the 2026 todo list?

List of vibe coded alternatives:

- revdiff: https://github.com/umputun/revdiff

- hunk: https://github.com/modem-dev/hunk

- diffnav: https://github.com/dlvhdr/diffnav

- diffx: https://github.com/wong2/diffx

- lumen: https://github.com/jnsahaj/lumen

Incredible, I have never heard of std::autodiff before. Isn't it rare for a programming language to provide AD within the standard library? Even Julia doesn't have it built-in, I wouldn't expect Rust out of all languages to experiment it in std.

First, you don't have to feel guilty of anything, since forking open source projects to make changes tailored to your use case is as old as open source itself. It is, in fact, the primary benefit of open source.

Second, it is not a given that your change would be accepted regardless of who wrote it. Maybe the feature is too niche for its complexity, maybe it is better implemented with more generality or extensibility that does not make sense for your own use. In those cases, your change might have been rejected upstream, so having it only locally is a perfect fine solution.

Third, if you believe it is actually useful for broader users, open an issue requesting that feature, and say LLM implemented it in an hour. Then the maintainers can prompt their own LLM to implement it with ease, or do whatever they want with their project.

Apple has device attestation deployed like one year before Google even proposed it: https://httptoolkit.com/blog/apple-private-access-tokens-att...

You want Typst: https://github.com/typst/typst

It's like the JSX of Latex: markup in a programming language, not a programming language pretends to be markup.

> learn SQL, even though it's out of fashion

In what world is SQL out of fashion??

Damn, we get it, USA is a dystopia. No need to keep scaring us with those stories.

They do have their own: https://hal.science/

It is actually quite common to come across HAL in subfields of mathematics in my experience.