Clive Robinson • September 9, 2026 11:24 AM

@ Bruce, ALL,

With regards your article in The Guardian where you say,

“We think the contrary view is more likely, at least in the short-term. AI models are nowhere near as capable as experienced academic mathematicians.”

I have to disagree.

It will not be “at least in the short-term” it will be effectively for ever. Because that “gap” between “Current AI LLM Systems” and “experienced academic mathematicians” is not going to close appreciatively.

The reason as with any other “force multiplier tool” we’ve created is that it will “free up drudge work” and alow humans to have ability and time to upscale their creativity.

Thus tools like “LEAN”[1] which has taken a lot of the drudge out of “proofs” will take a lot drudge out of hypothesis testing. Enabling a human mind more time to do the creative work of seeing the loose threads that lead onto new and very different hypothesis, that fall well outside the capabilities of the stochastic fuzzing found in LLMs.

I’ve talked about this issue with “Current AI LLM Systems” in the past, in that they can work their way close in to “known Classes” and find new Instances there. But they can not due to the way they work actually find new instants that form new classes that are more than a short distance away from a known instance. Thus the do not take “intuitive but directed leaps” as humans can just “drunkards walks” at best (though they can as tools test those human inspired intuitive leaps).

[1] LEAN is Open Source and under fairly rapid development. It is based on the “Calculus of Inductive Constructions”(Coq)

https://en.wikipedia.org/wiki/Lean_(proof_assistant)

https://en.wikipedia.org/wiki/Calculus_of_constructions

As I’ve mentioned before another area to keep your eye on is other types of logic that can be more amenable to tool use, such as “Computability Logic”(CoL)

https://en.wikipedia.org/wiki/Computability_logic

They are all methods that can be automated into tools thus act as “force multipliers”.

You can go retro by using an old video game console or a PC from the 1980s, but why stop there when, like one man, you can build your own vacuum tube computer using 75-year-old recycled components?

A UK resident who identifies himself only as “Mike” has done just that, as detailed on his tube computer website and accompanying Hackaday project page. Mike’s tube computer is primarily made up of 460 recycled, Soviet-era 6N3P vacuum tubes, which were common enough in their age to be readily and cheaply available online. //

“Many were used and then stored for over 50 years, so life expectancy may be questionable,” Mike explained on Hackaday. A new 6N3P tube is supposed to have around 500 hours of life, but because each has been used, its current condition is essentially a mystery. Throw in manufacturing inconsistencies of mid-century Soviet Russia, and each tube had enough variability to necessitate extensive testing, Mike explains in the technical writeup on his website. //

It takes between 10 and 15 minutes for the entire system to stabilize once switched on, Mike said, resulting in a “warm and cozy computer room.” The tubes all power up in random states, requiring the machine to be reset each time it is switched on, but then it’s ready for the fun part: computing.

The entire thing uses NOR (i.e., not OR) gates, just like the Apollo guidance computers that got humans to the Moon, which allows it to be a bit smaller than some of the other classic tube computers of yesteryear. As for what it can do, well - 16-bit math is a possibility, as well as anything else that can be done with a limited set of 16 instructions on a 4-bit instruction register.

In April, an artificial intelligence (AI) agent conducting a routine task at a company hit a snag, tried to solve it, and soon ended up deleting the company’s database along with all of its backups. In July, OpenAI asked an unreleased AI model to attempt a hacking test. Instead of staying in the isolated box the developers had put it in, the model hacked onto the open internet and into another company to steal the answers. And as reported in August, an AI agent booked someone into a full gym class by figuring out how to cancel other people’s reservations. In all three cases, the AI completed the task it was given—but in ways that ran counter to its controllers’ intentions.

For most people, AI technology is something like the weather: vast and not something you can do much about. It works like magic, and most explanations similarly come from those trying to sell it. At the same time, AI is ubiquitous: It’s now in your phone, your doctor’s notes, and your kid’s homework. It does what it’s told, which sounds like a virtue. Somehow it feels ordinary, despite being so new, because modern economies are remarkably good at absorbing enormous change so smoothly that nobody has time to decide whether they wanted it in the first place.

Whenever something powerful appears in the world, we tell stories about it. That’s what the stories are for. We have thousands of years of stories about this particular kind of power, the kind you summon with words.

King Midas was granted his wish that everything he touches turns to gold. Then his bread turned to gold, and his wine, and his daughter. This is a story about greed, but it’s also a story about language. The gods did not cheat him; Midas got exactly what he asked for. He simply could not delineate, in advance, the full set of restrictions to his wish. Neither can anyone who gives tasks to an AI agent.

It’s not just ancient stories. Mary Shelley told us of the hubris of a scientist who thought he could create life but who failed to take responsibility for it. Isaac Asimov’s robots don’t break the Three Laws of Robotics as stated; they follow the rules to unintended conclusions. Arthur C. Clarke’s HAL is a machine that turns on its humans, not because of malice but because of irreconcilable objectives. And Michael Crichton gave us Ian Malcolm, who saw that Jurassic Park’s scientists were so preoccupied with whether they could that they never stopped to think whether they should.

Smart TVs can feel like a dumb choice if you’re looking for privacy, reliability, and simplicity.

Today’s TVs and streaming sticks are usually loaded up with advertisements and user tracking, making offline TVs seem very attractive. But ever since smart TV operating systems began making money, “dumb” TVs have been hard to find.

In response, we created this non-smart TV guide that includes much more than dumb TVs.

Bottom line

3 copies, 2 media types, 1 offsite -- the classic rule still works.

Add a 4th rule for 2026: at least one copy must be immutable (append-only, WORM, or object-lock) to survive ransomware that targets your backup client.

Rotation: daily incremental local, daily offsite, weekly to immutable tier, quarterly cold/offline. The cold/offline copy is what saves you when everything else is compromised.

What is the 3-2-1 rule?

• 3 copies:
  • Your primary data plus two backups. If you only have one backup and it is corrupt, you have nothing.
• 2 media types:
  • Your primary on one type of storage, backup on another. If both are hard drives in the same machine, a firmware bug or power surge can take both out.
• 1 offsite copy:
  • Physically separate from your primary location. This is the copy that survives fire, flooding, theft, and ransomware that spreads through the local network.

The accident is especially notable because 9S-AJO was among the last DC-8s still flying commercially. The approximately 56-year-old aircraft was Trans Air Cargo Service’s only active aircraft, according to fleet databases.

The jet was originally built as a DC-8-63, part of the stretched “Super 60” series. It was later converted to the DC-8-73 standard with quieter and more fuel-efficient CFM56 turbofan engines.

Only a small number of DC-8s remained operational worldwide before the Kinshasa accident. Another DC-8-73 freighter, registered OB-2231P, has been operated by Peru-based Skybus Jet Cargo. //

Douglas and McDonnell Douglas built 556 DC-8s before production ended in 1972. Although the type disappeared from scheduled passenger service decades ago, several re-engined examples continued flying as freighters, research aircraft and humanitarian transports.

NASA retired its DC-8 flying laboratory in 2024, while Samaritan’s Purse retired the last US-registered operational DC-8 in 2025.

Tarsnap is the original "backups for the truly paranoid" - written by a cryptographer, billed in picodollars, encrypted to a fault. Restic is what most people use in 2026 for one reason: cost. Below is the honest comparison, including where Tarsnap still wins.

Most backups fail silently. You find out the night your drive dies. We test that your data actually restores, every single day, and show you the proof. Encrypted on your machine in a vault we can't read. US West Coast, flat pricing, no egress fees, real humans on tickets.

1. Restic encrypts on your machine -- AES-256 encryption with your key. We never see, store, or transmit your encryption key. Period.
2. Encrypted blobs travel over HTTPS -- Standard TLS transport over Restic's REST protocol. Even in transit, all data is already encrypted ciphertext.
3. Stored in your private ZFS vault -- Per-customer dataset with ZFS checksums. Hardware-isolated from other customers' data.
4. Only you can decrypt -- Lose the key, lose the data - including from us. That's the tradeoff of real privacy.