That's very helpful, thanks! The KOreader UI is... Rough.
Love all the options, especially the ability to use folders on the device storage as the primary way to organize my books, but trying to figure out where a setting is and change it has been frustrating from the first day I installed it.
Will give the Zen UI plugin a go. Does it still support the on-device folders as the main way to navigate through your books?
> All the AI companies are retaining the messages we're sending to their agents to train the next models. I wouldn't be surprised if the agents will be able to do what I'm doing in a year or two.
Some things are very easy to solve via code (see https://xkcd.com/1425/), and others are extremely difficult.
For LLM based AI, it can be trained to write functionally correct code easily - there's tons of pre-existing code and documentation that can let AI learn patterns (LAMP stack, messaging queue, etc) and syntax.
Judgement is much more difficult. For some things "Should I go with a message queue for this problem?", there's a known set of solutions as to what works best in a specific situation, sure, LLMs can answer that easily.
But when it comes to turning human goals into code that does the job correctly and is maintainable and stable?
That's a heck of a lot different. Without human guidance and having tons of context gained by years of experience dealing with the very nuanced and situational judgement calls needed, LLMs are going to fail at doing a decent job.
LLMs (and junior people) don't know what they don't know. I think it will be really difficult to get a LLM to work well at a senior / staff level where it's more about coordination, working with human beings, handling situations based on legacy code, knowing what will and won't be fragile, etc.
> I found it interesting that not even this article thinks about it that the US mainland ever can be attacked
It can be, but it would be very, very difficult for anything short of lobbing ICBMs around. You'd have to have a fleet of ships that would be detected as soon as they set sail, and then protect that fleet for the entire voyage, which would also be extremely difficult for any adversary.
Getting boots on US soil would be even tougher.
Drones are an option, but cross-ocean ones are not an easy problem to solve.
Getting them there seems easy, its the keeping them there that seems like a logistical nightmare.
>You'd have to have a fleet of ships that would be detected as soon as they set sail, and then protect that fleet for the entire voyage, which would also be extremely difficult for any adversary.
Or make friends with one of your neighbors that the USA appears to be keen on pissing off constantly. I have never seen more negative sentiment about the USA from Canadians before, who now see the USA as a strategic threat instead of their mentally challenged neighbor.
And dont get me started on Mexico, they can probably be had for pennies.
According to Zelensky, there will be drones capable of traversing tens of thousands of kilometers within a couple of years. The US mainland is definitely at risk of drone attack in the near-medium term, IMO.
> One, two, three years — drones will strike ten, twenty thousand kilometers. With reactive engines, they will be very cheap -- Zelensky
I did some digging, reaction engines are jet engines that can transition into a rocket propelled mode. It could enable a single stage to orbit plane.
Within the atmosphere the jet engine will burn atmospheric oxygen which would theoretically reduce the amount of fuel needed to reach a certain delta-V. Once outside of the atmosphere the engine switches to a rocket propelled mode, usually by injecting stored oxygen into the engine in addition to jet fuel.
So far there hasn't been a working prototype of a reaction engine. A British company SABRE closed down due to lack of funding but it did prove several pieces of the reactive engine design could work independently.
I have a friend whose dad finished his engineering career at SABRE (he retired some years before it shut down). According to him, the original concept wasn't far-fetched, but that they didn't crack some manufacturing / materials issues. He thought they should have continued to work on those, instead of redesigning to avoid them. He had retired by the time we spoke about it, and I don't know the particular model / design to which he was referring. If that's accurate, then possibly materials technology / manufacturing techniques have matured to the point that the "simpler" design he preferred could be feasible. That's all second-hand speculation, so take it with a dose of salt, but a short timeline might be reasonable.
Balloon-lofted drones seem more likely as a means for even a minor military power to strike deep within a country.
These can be launched in mass, cheaply. They're difficult to detect and/or distinguish (weather balloons are already common), and easy to cloud with dummy balloons, perhaps dangling aluminium paper to enhance radar signatures.
The drones themselves could use optical+inertial guidance, and have either minimal propulsion or be entirely glide-based. Given prevailing high-altitude winds (generally west-to-east in mid-northern latitudes) strikes could be launched many thousands of miles from targets with minimal or no fuel / propulsive storage use.
North Korea and Iran both have orbital launch vehicles. If you can put something in orbit, that's most of the way to putting something in Times Square or DC.
This hasn't happened yet. And much easier, more deniable attacks like car bombings of these places also hasn't happened to any real degree.
The mass shootings in the US are mostly performed by Americans, usually right-wing Americans, and the percentage performed by foreigners is close to zero. Looking forward to some right wing American replying linking to individual examples of foreigners doing mass shootings as though that disproves the point.
It could be attacked easily via an alliance with Mexico or Canada. The premise that such a thing could never happen is not a good one for military planning.
That does not sound stupid, but safe? I giant truck that has no chance of stopping, controlled by a computer? Just build railways and then there are no issues, no fancy AI to control them.
You misunderstand the mechanism. Its not because states copy other states. Its because the corrosive political elements are embedded everywhere, but most prominently have a foothold in California, Washington, and New York state. There is an interstate conspiracy and agenda to ruin America. Louis Rossman covered it recently in his video "The destruction of 3D printing: Bloomberg is behind it"
We abrogated getting traffic to our websites to Google long ago. Mostly because Google was so good at it that the alternatives became significantly less useful.
Now that Google is focusing on becoming 'self contained', so to speak, we should find a better way to drive traffic to websites. Ideally one that's not under the control of a single corporation.
I know this is likely to do with the nature of the problem, but that hasn’t stopped us from getting some wildly-unsuitable decentralised nonsense in the past.
It is hard to replace the old thing with another old thing. Search engine is already a fine tuned business, new comers will have hard time in it, no matter their tech stack.
It's like trying to raise better horses, while the other side has already built a empire on that and weaponized it.
The way out of here is to find something better than search engine, just like how cars replaced horses. But it's the same reason Google Search is replacing itself with AI too, they're already trying to replace their horses with cars.
On a higher level than individual URLs and separate from browser favorites. Something like versioned packages of links with decorations.
Something like a ".urlpackage" format that will have
- a list of urls
- optional metadata for each url, such as image, description, last-known-good
- metadata for the entire package, including version, an image, a favicon, and a description for the entire package that a client could use to present it nicely to the end user.
It'd be cool if my phone could open this format, show me the image and description with the list of links, and let me browse them, add them to my bookmarks, or add to the collection and make a new .urlpackage that I could then share back or publish somewhere.
It's probably possible to simply do this with a self-contained HTML file or similar I guess, though.
Does a move like this give more power / value to websites like reddit? A link aggregator that is organized is much more useful for finding new websites.
AI has made my work about 5-8x quicker, just because I'm able to have it cover a lot of the grunt work (update 42 if statements in 32 different files) that took time, but no particular skill.
I think the use cases where AI makes an economic improvement to the status quo for a business are rare, but they do exist, and they can be a significant improvement.
It's like the early days of the dotcom boom and bust - people thought the internet was good for every use case under the sun, including shipping people a single candy bar at a loss. After the dotcom bust, a lot of that went by the wayside, but there was a tremendous economic advantage to the businesses that were more useful when available on the internet.
is a silly behavior for a programmer or an AI to have to do more than twice. We have tools that very effectively remove the need for things like that: programming languages that allow modular and reusable code, good design, etc.
Ideally. But that requires the correct abstraction, requires keeping it up to date.... that's basically an unachievable ideal. You either have overabstraction/overengineering (most codebases) or you have repetition. Repetition is actually more preferable in the LLM-world because you have to keep less stuff in your head. And the LLM's head too.
Even if something does look copypasted, it might actually be semantically distinct enough that if you couple them, you'll create a brittle mess.
Additionally, there's always going to be global changes (update the code style, document things, refactor into a new pattern, add new functionality to callers, etc.). The question isn't whether you use your lanuage's tools or you do it by hand, the question is whether you use an LLM or do it by hand :P
Totally fair, but 42 if-statements across 32 files isn't something you need to fix with like ... a grand refactor or hexagonal architecture or event sourcing or whatever the overengineering pattern du jour is. You can fix that with a utility function or three, and a file/class/module/whatever that owns the code relating to some of those conditions.
I'm not some DRY zealot, but I've been in the "this system needs really similar changes to a ton of geographically distant code for simple changes" salt mines a lot. The people who say that kind of spaghetti is unavoidable are just as wrong as the ones who say it can only be fixed with a grand rearchitecture by a rockstar.
Sure but even wiring that utility function in is work :D If you have even just a 2-3-million LoC codebase, not even something truly enormous - making global changes does require typing, and a whole lot of it...
Such repetitions can regularly be deterministically automated, like find -exec sed and similar medium level tools.
If you spend a lot of time performing monotonic tasks, then your organisation needs to delete and refactor for a while until change in 'hot' areas of the code base are easy to make. Reaching for some code synthesis SaaS to paper it over will worsen the problem and should result in excommunication from the guild.
If you have a codebase that big, can you even fit enough of it into a context window for the LLM to make correct and meaningful changes across all of it? Admittedly I've only used LLM-based coding for smaller projects.
All of it hell no :D But just with any things, you break things down into subtasks. Then you break it down even more. You as a human don't hold all that stuff in your head either, so why would an LLM?
My current codebase is ~3 million LoC all in all (not greenfield, really old code), working on it by myself, the complexity is definitely manageable between Claude and me :)
I just want to mention that in my personal anecdotal experience, every codebase I have ever worked on, except 1, was underengineered and not overengineered. The last one was just "engineered".
Could you please show us an example of the change made to one of these if statements? I'm curious, because it seems absolutely wild to me to end up in such a situation (where that many changes are required and the usual refactoring tools of modern IDEs are insufficient) in the first place.
Does your work primarily consist of updating 42 if statements in 32 different files? We all do that occasionally, but if you're doing it constantly, is it possible that a different system design would make your work much easier?
If you are 8x quicker by having the AI do these for you, I think you are a junior intern or something? It must mean most of your time is spent doing these things.
Love all the options, especially the ability to use folders on the device storage as the primary way to organize my books, but trying to figure out where a setting is and change it has been frustrating from the first day I installed it.
Will give the Zen UI plugin a go. Does it still support the on-device folders as the main way to navigate through your books?