Hasnain says:

Lots of valuable quotes here (especially around the judicious doubt and worry about LLM abuse). But the author’s perspectives seem to match mine. I’ll leave with the conclusion:

“One of the most common retorts I get after showing these examples is some form of the statement “but those tasks are easy! Any computer science undergrad could have learned to do that!” And you know what? That's right. An undergrad could, with a few hours of searching around, have told me how to properly diagnose that CUDA error and which packages I could reinstall. An undergrad could, with a few hours of work, have rewritten that program in C. An undergrad could, with a hours hours of work, have studied the relevant textbooks and taught me whatever I wanted to know about that subject. Unfortunately, I don't have that magical undergrad who will drop everything and answer any question I have. But I do have the language model. And so sure; language models are not yet good enough that they can solve the interesting parts of my job as a programmer. And current models can only solve the easy tasks.

But five years ago, the best an LLM could do was write a plausibly-English sounding paragraph. And we were amazed when they could form coherent ideas from one sentence to the next. Their practical utility was exactly zero. Today, though, they've improved my productivity at the programming aspects of my job by at least 50% on the average project, and have removed enough of the drudgery that I built several things I would never have attempted otherwise.”

Hasnain says:

Looking forward to reading the full detailed analysis of the 200+ page ruling when they come out, because this is one for the history books.

"One of the most significant revelations from the case was the size of Google’s payments to Apple to secure the default search engine spot on iPhone browsers. An expert witness for Google let slip that the company shares 36 percent of search ad revenue from Safari with Apple. In 2022, Google paid Apple $20 billion for the default position.

During closing arguments, Mehta homed in on those payments, wondering how other players in the market could possibly displace Google from that position. “If that’s what it takes for somebody to dislodge Google as the default search engine, wouldn’t the folks that wrote the Sherman Act be concerned about it?”"

Hasnain says:

“Tiny code contests have historically focused on the source code. What makes that problematic is how it incentivizes obfuscation. On the other hand, optimizing for binary size usually results in the source code becoming more elegant. In order to make program files tinier, one must choose data structures and design patterns that are harmonious with the way computers work. Doing that requires us to better understand the nature of the problem.”

Hasnain says:

Can confirm.

“"It is fascinating to see historical figures’ legacy records convey crucial scientific implications to modern scientists even centuries later," said co-author Sabrina Bechet of the Royal Observatory of Belgium. "I doubt if they could have imagined their records would benefit the scientific community much later, well after their deaths. We still have a lot to learn from these historical figures, apart from the history of science itself. In the case of Kepler, we are standing on the shoulders of a scientific giant."”

Hasnain says:

“"It is fascinating to see historical figures’ legacy records convey crucial scientific implications to modern scientists even centuries later," said co-author Sabrina Bechet of the Royal Observatory of Belgium. "I doubt if they could have imagined their records would benefit the scientific community much later, well after their deaths. We still have a lot to learn from these historical figures, apart from the history of science itself. In the case of Kepler, we are standing on the shoulders of a scientific giant."”

Hasnain says:

“Why was Justin able to do this? Because this codebase had no master plan. There was no overarching design the system had to fit into. No expected format for APIs. No documented design system. No architectural review board making sure things were coherent. The app was a complete and utter mess. No one could ever fix it, so no one tried to. What did we do instead? We carved out our own little world of sanity.

This monolithic app, due to sheer necessity, had grown to be a microcosm of nice, small apps around its edges. Each person, when tasked with improving some part of that app, would inevitably give up untangling that web, and find some nice little corner to build new things. And then slowly update links to point to their nice new stuff, orphaning the old.

This may sound like a mess to you. But it was remarkably enjoyable to work in. Gone were the concerns of code duplication. Gone were the concerns of consistency. Gone were the concerns of extensibility. Code was written to serve a use, to touch as little of the area around it as possible, and to be easily replaceable. Our code was decoupled, because coupling it was simply harder.”

Hasnain says:

Isn’t this voter fraud?

“If they agree to submit all that, the system still does not steer them to a voter registration page. Instead, it shows them a “thank you” page.

So that person who wanted help registering to vote? In the end, they got no help at all registering. But they did hand over priceless personal data to a political operation.”

Hasnain says:

“That process involved several surprising steps, including “what’s in my mind the most important idea, which still seems a bit magical to me,” Maynard said. At one point, there was a seemingly obvious step they should have taken to simplify their sum. Instead, they left it in its longer and more complicated form. “We do something that at first sight looks completely stupid. We just refuse to do the standard simplification,” Maynard said. “And this gives up a lot. It means that now we can’t get any easy bound for this sum.”

But in the long run, this turned out to be an advantageous move. “In chess you call it a gambit, where you sacrifice a piece to get a better position on the board,” Maynard said. Guth likened it to playing with a Rubik’s Cube; sometimes you have to undo previous moves and make everything look worse before finding a way to get more colors in the right place.”

Hasnain says:

“The Plague Tale games are a fun and fairly unique mix of survival horror, puzzle solving, and stealth gameplay that is all reliant on the core tech of the not-so-intelligent rats. The same system as described throughout this video was adopted in both A Plague Tale Innocence and Requiem, with the rats in the sequel being more deadly. The rats at better at reacting to light, more likely to attack on the fringes of light sources and can even climb up walls or operate on multiple floors. But as stated, the most impressive element is how they're optimised, now with 300,000 rats being able to pathfind at once. And it's through one simple yet effective piece of AI tech, that the game’s design goals are realised.”

Hasnain says:

I am glad he’s dead but also not looking forward to the regional war this might start by the time I wake up..