Quotes

Lord Clement-Jones: To ask His Majesty's Government what assessment they have made of the cybersecurity risks posed by prompt injection attacks to the processing by generative artificial intelligence of material provided from outside government, and whether any such attacks have been detected thus far. Lord Vallance of Balham: Security is central to HMG's Generative AI Framework, which was published in January this year and sets out principles for using generative AI safely and responsibly. The risks posed by prompt injection attacks, including from material provided outside of government, have been assessed as part of this framework and are continually reviewed. The published Generative AI Framework for HMG specifically includes Prompt Injection attacks, alongside other AI specific cyber risks.

We simply don’t know to defend against these attacks. We have zero agentic AI systems that are secure against these attacks. Any AI that is working in an adversarial environment—and by this I mean that it may encounter untrusted training data or input—is vulnerable to prompt injection. It’s an existential problem that, near as I can tell, most people developing these technologies are just pretending isn’t there.

Large teams spend more time dealing with coordination and are more likely to reach for architecture and abstractions that they hope will reduce coordination costs, aka if I architect this well enough I don’t have to speak to my colleagues. Microservices, event buses, and schema free databases are all examples of attempts to architect our way around coordination. A decade in we’ve learned that these patterns raise the cost of reasoning about a system, during onboarding, during design, and during incidents and outages.

A man sets himself the task of portraying the world. Through the years he peoples a space with images of provinces, kingdoms, mountains, bays, ships, islands, fishes, rooms, instruments, stars, horses, and people. Shortly before his death, he discovers that that patient labyrinth of lines traces the image of his face.

The cognitive debt of LLM-laden coding extends beyond disengagement of our craft. We’ve all heard the stories. Hyped up, vibed up, slop-jockeys with attention spans shorter than the framework-hopping JavaScript devs of the early 2010s, sling their sludge in pull requests and design docs, discouraging collaboration and disrupting teams. Code reviewing coworkers are rapidly losing their minds as they come to the crushing realization that they are now the first layer of quality control instead of one of the last. Asked to review; forced to pick apart. Calling out freshly added functions that are never called, hallucinated library additions, and obvious runtime or compilation errors. All while the author—who clearly only skimmed their “own” code—is taking no responsibility, going “whoopsie, Claude wrote that. Silly AI, ha-ha.”

Question to Radio Yerevan: “Is it correct that Grigori Grigorievich Grigoriev won a luxury car at the All-Union Championship in Moscow?” Radio Yerevan answered: “In principle, yes. But first of all it was not Grigori Grigorievich Grigoriev, but Vassili Vassilievich Vassiliev; second, it was not at the All-Union Championship in Moscow, but at a Collective Farm Sports Festival in Smolensk; third, it was not a car, but a bicycle; and fourth he didn’t win it, but rather it was stolen from him.”

For many, crypto had become an identity, a way to feel smart and subversive and on the cutting edge of a new technology. What happens to that self-image when its foundation erodes? When instead of being someone’s savvy son or daughter, you are the sheepish adult child who has to explain where the family savings went?

Let us now assume, for the sake of argument, that these machines are a genuine possibility, and look at the consequences of constructing them...it seems probable that once the machine thinking method had started, it would not take long to outstrip our feeble powers. There would be no question of the machines dying, and they would be able to converse with each other to sharpen their wits. At some stage therefore we should have to expect the machines to take control, in the way that is mentioned in Samuel Butler’s Erewhon.

In its light we must think and act not only for the moment but for our time. I am reminded of the story of the great French Marshal Lyautey, who once asked his gardener to plant a tree. The gardener objected that the tree was slow-growing and would not reach maturity for a hundred years. The Marshal replied, "In that case, there is no time to lose, plant it this afternoon."

The Platonists sense intuitively that ideas are realities; the Aristotelians, that they are generalizations; for the former, language is nothing but a system of arbitrary symbols; for the latter, it is a map of the universe...Across latitudes and epochs, the two immortal antagonists change languages and names: one is Parmenides, Plato, Spinoza, Kant; the other, Heraclitus, Aristotle, Locke, Hume, William James...Maurice de Wulf writes: "Ultra-realism garnered the first adherents. The chronicler Heriman (11th century) gives the name 'antiqui doctores' to those who teach dialectics in re; Abelard speaks of it as an 'antique doctrine', and until the end of the 12th century, the name moderni is applied to its adversaries." A hypothesis that is now inconceivable seemed obvious in the 9th century, and lasted in some form into the 14th. Nominalism, once the novelty of a few, encompasses everyone; its victory is so vast and fundamental that its name is useless. No one declares himself a nominalist because no one is anything else. Let us try to understand, nevertheless, that for the men of the Middle Ages the fundamental thing was not men but humanity, not individuals but the species, not the species but the genus, not the genera but god. From such concepts...allegorical literature, as I understand it, derived.

If nature has made any one thing less susceptible than all others of exclusive property, it is the action of the thinking power called an idea, which an individual may exclusively possess as long as he keeps it to himself; but the moment it is divulged, it forces itself into the possession of everyone, and the receiver cannot dispossess himself of it. Its peculiar character, too, is that no one possesses the less, because every other possesses the whole of it. He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me. That ideas should freely spread from one to another over the globe, for the moral and mutual instruction of man, and improvement of his condition, seems to have been peculiarly and benevolently designed by nature, when she made them, like fire, expansible over all space, without lessening their density in any point, and like the air in which we breathe, move, and have our physical being, incapable of confinement or exclusive appropriation. Inventions then cannot, in nature, be a subject of property.

For those who haven't heard the story the details were pulled from a Christian dating site db.singles.org which had a query parameter injection vulnerability. The vulnerability allowed you to navigate to a person's profile by entering the user id and skipping authentication. Once you got there the change password form had the passwords in plain text. Someone wrote a scraper and now the entire database is on Mediafire and contains thousands of email/password combinations.

50% of cybersecurity is endlessly explaining that consumer VPNs don’t address any real cybersecurity issues. They are basically only useful for bypassing geofences and making money telling people they need to buy a VPN. Man-in-the-middle attacks on Public WiFi networks haven't been a realistic threat in a decade. Almost all websites use encryption by default, and anything of value uses HSTS to prevent attackers from downgrading / disabling encryption. It's a non issue.

With enough users, every observable behavior becomes a dependency - regardless of what you promised. Someone is scraping your API, automating your quirks, caching your bugs. This creates a career-level insight: you can’t treat compatibility work as “maintenance” and new features as “real work.” Compatibility is product. Design your deprecations as migrations with time, tooling, and empathy. Most “API design” is actually “API retirement.”

If you had a choice between the ability to detect falsehood and the ability to discover truth, which one would you take? There was a time when I thought they were different ways of saying the same thing, but I no longer believe that. Most of my relatives, for example, are almost as good at seeing through subterfuge as they are at perpetrating it. I’m not at all sure, though, that they care much about truth. On the other hand, I’d always felt there was something noble, special, and honorable about seeking truth... Had this made me a sucker for truth’s opposite?

A King may move a man, a father may claim a son, but remember that even when those who move you be Kings, or men of power, your soul is in your keeping alone. When you stand before God, you cannot say, "But I was told by others to do thus." Or that, "Virtue was not convenient at the time." This will not suffice. Remember that.

When human beings found out about death They sent the dog to Chukwu with a message: They wanted to be let back to the house of life. They didn't want to end up lost forever Like burnt wood disappearing into smoke And ashes that get blown away to nothing. Instead, they saw their souls in a flock at twilight Cawing and headed back for the same old roosts (The dog was meant to tell all this to Chukwu)...

Ask browser users, and they'll tell you the overwhelming reason why they can't upgrade to a more modern, standards-compliant browser is because their work won't let them. Ask IT departments why this is the case and they'll point to the six- to seven-figure costs of upgrading turn-of-the-century Intranets written to work in, and only in, Internet Explorer 6. Google have provided a way for websites to opt out of IE6 (and even IE7) support without requiring enterprise-wide, Intranet-breaking browser upgrades.

The internet is self destructing paper. A place where anything written is soon destroyed by rapacious competition and the only preservation is to forever copy writing from sheet to sheet faster than they can burn. If it's worth writing, it's worth keeping. If it can be kept, it might be worth writing. Would your store your brain in a startup company's vat? If you store your writing on a third-party site like Blogger, LiveJournal or even on your own site, but in the complex format used by blog/wiki software de jour you will lose it forever as soon as hypersonic wings of internet labor flows direct people's energies elsewhere. For most information published on the internet, perhaps that is not a moment too soon, but how can the muse of originality soar when immolating transience brushes every feather?

In the Bay Area, we have a collection of fiefdoms. Villages are parading as cities, addressing problems myopically. For example, Brisbane (a city of 5,000 people between San Francisco and SFO) is currently blocking a large housing development for local reasons. It’s NIMBY-ism on a broad scale – a regional tragedy of the commons.

Serverless is a somewhat unhelpfully misleading term for "highly scalable stateless code". All the times I've seen serverless stuff work really well it was workloads that were usually zero but occasionally 30k/sec without warning. I've run a company with that kind of workload and serverless stuff would have saved us a ton of money. Publishing to the [npm] registry could be done as a serverless app but there's little benefit because we do not get huge spikes in publishing. We get huge spikes in downloads but serverless isn't useful there because it's a read-only case and very little processing is done. Serverless is a great solution to one type of problem. It's very seldom the case that you can convert all your problems into that shape.

For quite some I wanted to write a small static image gallery so I can share my pictures with friends and family. Of course there are a gazillion tools like this, but, well, sometimes I just want to roll my own. [...] I used the old, well tested technique I call brain coding, where you start with an empty vim buffer and type some code (Perl, HTML, CSS) until you're happy with the result. It helps to think a bit (aka use your brain) during this process.