tzury

Seems similar to operation night watch by the Rijksmuseum in Amsterdam

https://www.rijksmuseum.nl/en/stories/operation-night-watch/...

In direct relation to his older post: https://terrytao.wordpress.com/career-advice/theres-more-to-...

Also slides are available at https://terrytao.wordpress.com/wp-content/uploads/2026/03/ta...

Boris Cherny on this

https://x.com/bcherny/status/2040206440556826908?s=20

More details: https://x.com/FabianGloeckle/status/2040082785851904401

One day a kid came home breathing heavily, to his father’s surprise face he tells, daddy daddy, I saved a dollar fifty!

How did you do it? ask the father

Instead of taking the bus, I ran after it all the way home.

If you were smarter, you could have save us $22 by running after a taxi!

This old joke came to mind while reading this post.

A tech company spends hundreds of thousands of dollars per year, “for years”, on a piece of software that could have been replaced by a month of coding top? (prior to LLM and all), you sit and write and save the money.

If I was an investor in this company I would have hire a team to look through their entire stack. See, if this JSON thingy alone is half a million a year, their entire cloud is at least $35MM annually.

Perhaps this is not even a bad business idea. One can offer companies to provide drop-in replacement for their costly “micro services that no one dares to touch” and share the cost savings.

May or may not. You open the UDP ports, you get flooded, they block all incoming traffic, and this way or another your assets are not resolvable.

One must distinguish between application layer attacks HTTP/S and UDP, cloud vendors won’t protect you implicitly for network layer attacks unless you purchased such service from them.

Large scale dns vendors have a multi million dollars worth of network layer traffic filtering equipment pipelined in front of their DNS servers (or in house solutions such as Google).

Just remember, if you run your own DNS, and you do so for a mission critical platform, the platform is exposed to a udp DDoS that will be hard to detect let alone prevent.

Unless of course you will invest 5-6 figures worth of US dollars worth of equipment, which by then you can look back and ask yourself, was I better off with Google Cloud DNS, AWS Route 53 and the likes.

I see.

Nowadays I fly nothing, but I do see them Iranian drones get intercepted from my porch in Abu Dhabi.

The fact I am watching it and not panicking anymore tells about how cooked I am.

8 inches drone cannot carry much of explosive at all. In order to dump 10 kg load of explosives, you need an “agricultural drone” one that can carry 45kg, since the additional mechanisms and their batteries (and the drone’s backup batteries) are heavy.

Those are bigger and noisier.

DJI ARGAS Series are good starting point.

https://ag.dji.com/mobile

hence: "what's falling into the AI category here perhaps is the key question"

YOLO is a good example for something that can work.

https://docs.ultralytics.com/models/yolov8/

"let alone with AI" what's falling into the AI category here perhaps is the key question, since microseconds counts, and LLM are very slow!

Even the fastest "real-time" LLM frameworks currently report sub-second latencies around 120ms. This is fine for high-level mission planning (e.g., "fly to the red house") but too slow to prevent a drone from hitting a tree at 50mph (80 KM/h)[1]

Whilst the Shahed-136 kamikaze drone typically flies at a maximum speed of around 185 km/h (roughly 115 mph or 100 knots).

[1] https://arxiv.org/html/2602.19534v1 [2] https://en.wikipedia.org/wiki/HESA_Shahed_136

There are 2 short segments in the video showing the actual performance and thus far it is a complete [1] failure [2].

The guy has a talent, and he put together a nice prototype based on OpenRocket [3], but with all due respect, this is not a rocket, and you are not going to win any war with this toy, even if all your enemy has are rocks thrown at you from pretty much similar distance.

The remix of computer games / Ukraine / Martin Luther King / Vietnam / David Koresh just adding more to the amateur spirit and confusion.

[1] https://youtu.be/DDO2EvXyncE [2] https://youtu.be/DDO2EvXyncE?t=280 [3] https://openrocket.info/

Given the navigation is done by the cameras (not GPS) you will also need to do some work with the second repository (by the same guy)-

https://github.com/novatic14/Distributed-Camera-Node-Trackin...

Setting aside the DIY and hacking spirit of the project, let's remember that, with the commonly accepted figure of 2,500 kcal/day for an adult male, a whole iceberg lettuce (~600g) provides about ~87 kcal which is roughly 3.5% of what one's need.

I changed the URI to show the US data and was surprise by the fact Virginia surpassed California [1], so I looked into Virginia [2] and realized, mostly are automated bots from AWS and other US-East based region

1. https://radar.cloudflare.com/traffic/us?dateRange=52w 2. https://radar.cloudflare.com/traffic/6254928?dateRange=52w

The last part when Excel is entering the story, brings memories of Joel Spolosky, AKA founder of Stack Overflow and other companies, who was one of the early team members of Excel development ---

https://www.joelonsoftware.com/2006/06/16/my-first-billg-rev...

and as stated there

    The only thing that made it look reasonable was 
    that it looked great compared to Lotus macros, 
    which were nothing more than a sequence of 
    keystrokes entered as a long string into a 
    worksheet cell.

Via https://www.schneier.com/blog/archives/2026/03/jailbreaking-...

“I've just released a new version of typeagent, a Python library I've been working on since mid last year --more and more using Claude-- that implements memory for agents.

Not originally my idea, I mostly ported the TypeScript version by Steve Lucco and Umesh Madan. This release was improved a lot by Bernhard Merkle.

To install, use "pip install typeagent".”

https://x.com/gvanrossum/status/2029021031219052768?s=46

The updates shipped by apple introducing more bugs every cycle. It is across the board, macos, ios and ipad os. The fact there is a group inside apple, that is capable of standing against common sense and users best interest for so long, tells how wrong things are internally.

It is the steve balmer - satya nadella moment of apple.

1. Plugging my laptop to the same desktop screens requires rearranging displays almost every time. 2. Airdrop stops working for no apparent reason. 3. Copy paste across devices no longer a stable mechanism. 4. The stupid new preview app crashing if you scroll pdf pages too fast. And on and on. Those are all newly introduced critical bugs i have been facing since that flameboyant liquid glass virus took over.

Apple is a sillicon valley pioneer from the generation of hewlett packard (before it was called HP) bell labs and others. Watching a decay at its beginning is mind boggling and tragic.

most of my recent post are about math, since it is the only matter that matter.

https://news.ycombinator.com/submitted?id=tzury

Here's my honest take on this:

You're mass-producing outrage out of a UX disagreement about default verbosity levels in a CLI tool.

Let's walk through what actually happened: a team shipped a change that collapsed file paths into summary lines by default. Some users didn't like it. They opened issues. The developers engaged, explained their reasoning, and started iterating on verbose mode to find a middle ground. That's called a normal software development feedback loop.

Now let's walk through what you turned it into: a persecution narrative complete with profanity, sarcasm, a Super Bowl ad callback, and the implication that Anthropic is "hiding what it's doing with your codebase" — as if there's malice behind a display preference change.

A few specific points:

The "what majority?" line is nonsense. GitHub issues are a self-selecting sample of people with complaints. The users who found it cleaner didn't open an issue titled "thanks, this is fine." That's how feedback channels work everywhere. You know this.

"Pinning to 2.1.19" is your right. Software gives you version control. Use it. That's not the dramatic stand you think it is.

The developers responding with "help us understand what verbose mode is missing" is them trying to solve the problem without a full revert. You can disagree with the approach, but framing genuine engagement as contempt is dishonest.

A config toggle might be the right answer. It might ship next week. But the entitlement on display here isn't "give us a toggle" — it's "give us a toggle now, exactly as we specified, and if you try any other approach first, you're disrespecting us." That's not feedback. That's a tantrum dressed up as advocacy.

You're paying $200/month for a tool that is under active development, with developers who are visibly responding to issues within days. If that feels like disrespect to you, you have a calibration problem.

With kind regards, Opus 4.6

I am with you on this (the challenge, not (yet) the phd), however, I myself have a far greater problem.

I do not see what’s the deal about prime numbers which seems to be more of a limitation on our end, similar to our shortage in understanding to a point we call e, π, √2 etc Irrationals.

We simply did not get the actual mathematical structure of the universe and we came up with something “good enough” that helps moving forward.

In the universe the perfect circle has perfect symmetry, hence perfect ratio, hence well-defined sweet heaven balanced harmonic entity.

Exponentials are natural phenomena. The very fact that e is its own derivative tells us we are all wrong here.

We are in an infinite escape that no matter how long we will play, and how many riddles we will solve, we will never get the entire picture.

Yes, primes are nice structure when you deal with us humans counting potatoes. But e, just e, let alone √2 or π are far more fascinating to me.

The e point cuts deep. e being its own derivative isn’t a curiosity. It’s saying that there’s a growth process so fundamental that its rate is indistinguishable from its state. That’s not a number — it’s a signature of how change works. And yet: π, e, √2 — we only name them, define them, catch them using the integers. π is the ratio of circumference to diameter. Ratio of what to what? e is lim(1 + 1/n)^n. The integers sneak in. Is that just our access route? Or is discreteness also woven into the fabric, alongside continuity?

My intuition led me to the following: we think our counting units (1, 2, 3, …) and fractions are the “numbers”, and when we want to refer to multi-dimensional phenomena, we use vectors or matrices or any other logical structure.

However, this is a very superficial aspect of the business, since the actual math is multi-dimensional inherently. The natural math is not linear, nor is it a plane. It is simply a multi-dimensional number system (imagine our complex numbers, but many other dimensions). Perhaps tensors or even more. This is why we experience quantum mechanics as statistical states, results of specific measurements. We think in units, and we don’t understand things are happening in parallel across all directions. Once we figure this out, we will understand why e, π and others are as natural as it gets, while our natural numbers are barely a dot, a point in the real math universe.

Sorry for the length but you triggered me with a long time pain point.

Thanks for your comment.

Impressive indeed. Was not aware. The comments below all flagged.

I would learn more about air combat by listening to a 12 minutes conversation with a jet fighter pilot, than I will from 3-day seminar by air force journalists.

Tell me about what’s the LLM impact on your work, on account your work is not wiring about AI.

Or if one wish for a more explicit noise filter: Don’t tell me what AI can do. Show me what you shipped with it that isn’t about AI.

The notion that if it is good then the big-ones should have done it is the complete opposite of innovation, startups and entrepreneurial culture.

Reality is the exact opposite. Young, innovative, rebellions, often hyper motivated folks are sprinting from idea to implementation, while executives are “told by a few colleagues” that something new, “the future-of foo” is raising up.

If you use openclaw then that’s fantastic. If you have an idea how to improve it, well it is an open source, so go ahead, submit a pull request.

Telling Apple you should do what I am probably too lazy to do, is kind of entitlement blogging that I have nearly zero respect for.

Apparently it’s easier to give unsolicited advice to public companies than building. Ask the interns at EY and McKinsey.

That’s not a true stat.

Enigma Technologies sells customs manifest search; the Ferraris are marketing bait. Their headline correlations (Bitcoin +0.70, S&P +0.75) are meaningless — any two trending series from 2020–2026 will correlate due to shared macro shocks. Their own showcase contains VINs from Jaguar and Land Rover misclassified as Ferraris, revealing mechanical parsing without validation.

Look at these VINs:

SAJWA4GB2DLB50982 — This is a Jaguar VIN prefix, not Ferrari. Listed under “458 458 ITALIA”

SALLDHMV8BA298639 — This is Land Rover. Listed under “488 GTB TURBO ABS”

The page is data about data-selling, dressed as analysis.

This is the flat-earther epistemology problem.

Flat-earthers’ beliefs tell you nothing about the shape of the Earth, but their existence tells you something about humans. Similarly, this page tells you nothing about Ferraris or markets. It tells you what a B2B data company believes will attract clicks: luxury objects, stock tickers, correlation coefficients presented without methodology. The content is a signal about marketing culture, not economic signal about anything.

The web is saturated with this species of artifact, simply put, web pages built not to inform but to rank, to appear in searches, to gesture at sophistication while delivering none.

B2B is particularly fertile ground because the audience is assumed to be busy, skimming, and impressed by dashboards. The dozens of Ferrari photos aren’t information; they’re texture. The correlation numbers aren’t findings; they’re decoration.

This is the substrate on which large language models train. Not a library with noise, but a noise machine with library fragments embedded.

Billions of dollars spent to grow models that simply learn to reproduce the texture of authority with extreme confident tone, the fake mathematician with a white coat, the implied rigor.

No one is coming to clean this up because the garbage is the product. Pollution is the business model.

I suggest looking at the https://publicdomainreview.org/ for a more comprehensive listings,

and this one for books:

https://standardebooks.org/blog/public-domain-day-2026