I've had no success getting Agents to help write code for SQLite. All the code I've asked Claude to write for me is slop that didn't work or worked poorly and never got committed. On the flip side, I do ask Claude to review the code that I write prior to committing, and it is good at that, often finding serious oversights that I had missed.

This observation, that agents are good at code review but bad at writing original code, seems to carry through into AI-generated bug reports coming into SQLite from third parties. The bug reports are generally good, but the suggested fixes not so much.

One report from about a month ago illustrates this. The median() aggregate function in SQLite is implemented by collecting the input values into an array of doubles, using quicksort to put them in order, and return the one in the middle. The bug report showed a carefully devised sequence of 1 million inputs that caused pathological O(N*N) behavior in the quicksort algorithm, which caused recursion to go too deep and blow out the CPU stack. The AI's suggested fix: Add a depth parameter to the quicksort algorithm and fail over to a slower sorting algorithm if the recursion goes too deep. The correct solution (originally recognized in the 1975 by Robert Sedgewick, but unknown to me before this issue arose) was to only call quicksort recursively on the smaller of the two partitions, and sort the larger partition using a loop (tail recursion). If recursion only occurs on a partition that is half the size of the original or smaller, recursion depth is limited to LogN. That turned out to be trivial to implement, and is faster and uses less code than the AI-suggested approach. Furthermore, after seeing Sedgewick's tail recursion idea, Dan Kennedy recognized that we don't actually need to sort both partitions at each step of the algorithm, but only the partition that contains the median. Dan's observation more than doubled performance. Hence by ignoring the AI's suggested fix, we more than doubled the performance of the median() function in SQLite.

Before you ask: The median() function did originally use the standard library qsort(), but I changed that into a hand-coded quicksort some time ago because the hand-written variant did not require a callback for each comparison, and was thus way faster.

Even if China were to pass its AI tech to Russia, Russia no longer has the industrial base, nor the liquid assets to take advantage of it. And that's not counting the massive brain drain Russia experienced when almost 2 million men (many of them in tech) left at the start of the war.

Ukraine, OTH, hasn't suffered that type of brain drain and, thanks to Western support, it can carry a war economy and build AI-powered drone tech at scale.

This was a war for Russia to win if it hadn't become so incompetent and corrupt under Putin (and thank God for that).

Leadership matters.

It's not his mouth that he's opening, though.

News you don't like == propaganda?

Shit is happening on the tech front in the Russo-Ukrainian war with spillovers in the Middle East. We are witnessing a battleshift paradigm shift not unlike the widespread adoption of gunpowder, airpower and/or information technology....

... both sides have been heavily invested in this shit, not just Ukraine...

... and it's tilted the war in Ukraine's favor...

... and Hezbollah has embraced that 100%, giving a technically superior force (the IDF) more headaches than it can handle...

... and it's shaping how technology is going to be developed...

... and you think this shouldn't be covered in a technology-oriented website?

Oh my sweet summer child.

This reminds me of every single management moron that dismantled IT systems monitoring to save money because nothing happens anyway , ignoring that it's not that nothing happens but that the monitoring provides a way to see trends and take preemptive actions before SHTF.

Better, yet, it seems that projects that rely on paste.bin discussions should have been using a source control repository to keep them.

People need to treat discussions as the "live" part of technical documentation (which opens another can of worms, obviously.) But, if it is not source-controlled, assume it is ephemeral, or that it doesn't exist.

This also applies to stuff we cite in Stack Overflow. It's handy to cite a finding when documenting a code change, hack, or design decision based on an answer found in SO. But for very important stuff, additional documentation must exist somewhere, preferably source-controlled.

Why should I define it when there are already textbook definitions in psychology, neuroscience, philosophy, medicine, and AI research?

This is at the root of the problem. I don't know what Dawkins was referring to. The onus is/was on him to define it before allowing himself to be quoted this way.

And if I am pushed to define, I will decline. I will simply offer the mainstream definitions in the disciplines I've already mentioned.

This. Or buy it used.

Honestly, there's no excuse to pirate stuff over BitTorrent when one can buy a book in print.

No, it won't. Yes, there are ethical and copyright issues regarding LLM training data, but this is sabotage of something that is also quite useful to society.

We need to have a conversation about the ethical and social implications of AI. This ain't it. This is just edgelord shit hurling.

Richard Dawkins has said chatbots should be considered conscious (source paywalled; alternative source) after spending two days interacting with the Claude AI engine.

I can't believe someone like Dawkins would fall for anthropomorphizing AI chatbots... unless he's using a different definition of consciousness, which is fair.

So, we have to start there: what does "being conscious" mean, for this scenario, and for Dawkins while evaluating this scenario?

Seems like a rather subjective and emotionally charged perspective. Nothing wrong with that so long as we recognize (and he recognizes) it for what it is.

With that said, this is a conversation worth having... within certain parameters (tbd)

Words detected. Useful signal missing.

I tend to agree with the spirit of such comments, but in this case, Apple makes it clear in the TOS that it retains the right to disclose that information if it deems it appropriate. People pay for an anonymity service without reading the fine print :/

No. We saw a marked decrease in crime rates, starting around the 90s (a generation after the ban on leaded gas started)

For a country of Iran's size, its economy and public services can ill-afford this type of mass disruption - This is a "soft" Doomsday Device, except it is not soft at all.