post_below's recent activity
-
Comment on AI doomers: What uses of generative AI are you actually excited about? in ~tech
-
Comment on Building a C compiler with a team of parallel Claudes in ~tech
post_below Link ParentTo be fair to them, they didn't claim it was glorious in the blog post. It mentions that, with all optimizations enabled, it performs worse than GCC with all optimizations turned off. It also...To be fair to them, they didn't claim it was glorious in the blog post. It mentions that, with all optimizations enabled, it performs worse than GCC with all optimizations turned off. It also talks about the code quality being sub par.
The frustrating thing is that's not likely how the media and bloggers will talk about it. It will be another round of AI doom "it's coming for your job". It will fuel the hot takes that AI can now truly just write software. It will help suck in a new round of vibe coders. Except this year they want to be called "vibe engineers".
What it really is, like Cursor's far worse and more expensive example before it, is a somewhat interesting proof of concept. A few short years ago the possibility of agents creating any non trivial application autonomously was absurd.
I hate the hype too. But if the well wasn't poisoned by hype, and the airwaves saturated with AI discussion, we'd all be at least a little bemused by this.
-
Comment on Any software engineers considering a career switch due to AI? in ~comp
post_below LinkOption 2 is to build your own thing, which you can get started on any time, even keeping your current job and income. It's not for everyone but you'd get to decide exactly how much hands on...I really love building stuff and solving problems so maybe I go back to school and switch to some other flavor of engineering
Option 2 is to build your own thing, which you can get started on any time, even keeping your current job and income. It's not for everyone but you'd get to decide exactly how much hands on building and problem solving you'd get to do. The trick IMO is finding a problem you really care about solving, rather than solving a problem just to make money.
-
Comment on Passing question about LLMs and the Tech Singularity in ~tech
post_below Link ParentAh, yes in that context we're nowhere near an explosion. Or at least the existing technology doesn't put us near one, who knows if there will be breakthroughs in the near future. Yes LLMs are...Ah, yes in that context we're nowhere near an explosion. Or at least the existing technology doesn't put us near one, who knows if there will be breakthroughs in the near future.
Yes LLMs are already helping move the technology along faster than humans alone could do it. I don't think there's any doubt of that. The only question is if the path leads to the vicinity of AGI, which I think is safe to answer yes. It doesn't matter of LLMs themselves will have anything to do with AGI, they will definitely accelerate many aspects of technological advancement and some of them will contribute to eventual AGI.
-
Comment on The AI industry doesn’t take “no” for an answer in ~tech
post_below Link ParentI'd replace humans with mammals, and it's a well known part of the process of evolution. Calories are historically expensive and both movement and cognition use a lot of calories so organisms...I'd replace humans with mammals, and it's a well known part of the process of evolution. Calories are historically expensive and both movement and cognition use a lot of calories so organisms evolve to be as lazy as they can get away with while still surviving effectively.
-
Comment on Passing question about LLMs and the Tech Singularity in ~tech
post_below LinkAs others noted, LLMs don't exactly have "code". However, LLMs have been used as a part of the LLM training process for some time. In that sense, yes absolutely. In another sense: the agent...In other words, are we (the humans) already starting to use LLMs to improve their code faster than we humans alone could do?
As others noted, LLMs don't exactly have "code". However, LLMs have been used as a part of the LLM training process for some time. In that sense, yes absolutely.
In another sense: the agent harness, which is an increasingly important part of how effective LLM powered agents are at real world tasks, also yes. The big model companies use coding agents extensively when creating the harness and scaffolding, which shows in the sheer volume of bugs the harnesses have.
Wouldn't this be the actual start of the predicted "intelligence explosion"?
I didn't read the article you linked so I may not have all the nuance about what intelligence explosion refers to. But in terms of practical application of AI agents at real world tasks we are undoubtedly in the midst of an intelligence explosion.
In terms of actual intelligence, as we might define it, LLMs arguably have none at all. It depends on how you frame it. They provide an illusion of intelligence that is so good they can do intelligent things, which is to say things that could previously only be accomplished with human intelligence. But it's achieved, essentially, through advanced pattern matching rather than anything that could be described as understanding. It's hard to imagine a path from there to true AGI but at the same time it's not difficult to imagine that at some point the illusion of intelligence could get so good that it's practically the same as the real thing for many applications.
It's all so weird.
-
Comment on Why there's no European Google? in ~tech
post_below LinkI missed this thread when it was posted but my late entry is this: Putting the author's steriotypically French delivery aside, I think their point is fantastic. What if we looked at value...I missed this thread when it was posted but my late entry is this: Putting the author's steriotypically French delivery aside, I think their point is fantastic. What if we looked at value differently? We should do that.
What if Torvalds is more successful, as a human being, than Jeff Bezos? If the score isn't kept in dollars it's not a difficult case to make.
I know this seems naive to the point of absurd from an American perspective, where the market devours idealism without even trying, but from humanist perspective it's entirely valid.
-
Comment on Someone made a social media website for AI agents in ~tech
post_below Link ParentI doubt the goal of this project is training, more likely it's just a way to create buzz. However if it was about training, it's a dramatically cheaper way to create slop than paying for your own...I doubt the goal of this project is training, more likely it's just a way to create buzz. However if it was about training, it's a dramatically cheaper way to create slop than paying for your own tokens. You get hordes of other people to spend tokens for you.
-
Comment on New books aren’t worth reading in ~books
post_below LinkAlternative title: "New xitter posts are not worth reading" But I did read it, and now I'm here with my life and perspectives unimproved and unchanged. The post is clear ragebait, it's...Alternative title: "New xitter posts are not worth reading"
But I did read it, and now I'm here with my life and perspectives unimproved and unchanged. The post is clear ragebait, it's confrontational out of the gate, it just assumes rage. Maybe that's just how everyone talks on Xitter now, I don't go there often. It also establishes "this debate" out of thin air, though the nature of the debate is unclear. I assume they're using debate as a way to denote the rage they expect to illicit. Then at the end it just fizzles, having failed to pay off any of its claims.
I didn't read the replies because I don't want to log in but I can guess what they look like. With 300k views, the post accomplished its goal: Engagement with very little investement in time or mental energy on the author's part. The post is more interesting if you read it as a thinly veiled cry of frustration at the mental state of being overly online that the author is drowning in.
However, I enjoyed this bit:
The average ancient historian led troops, tutored a prince, governed a province, advised a king, made a fortune, fell from favor, was exiled, and buried 7 of their 10 children. The average modern historian passed a few tests then wrote a book on their laptop next to their cat.
It's kinda true! I disagree with most of the rest but that part changed the shape of my mouth.
-
Comment on Gold tops $4,900/oz; silver and platinum extend record‑setting rally in ~finance
post_below Link ParentCongrats on the success of your investments! Usually gold only behaves like this when there's a recession, and it's fantastic to get in early enough because there's so little risk that way....Congrats on the success of your investments! Usually gold only behaves like this when there's a recession, and it's fantastic to get in early enough because there's so little risk that way. Historically, while gold can drop in price, it usually stabilizes pretty quickly and, so far, always comes back at some point. If you buy into one of its upswings early enough it's very safe relative to the potential returns.
It's been wild the last couple years, pretty much every major financial institution has been low on projections and has revised their numbers upwards multiple times a year. Currently all the institutions that haven't already posted new predictions in January are due to upgrade their projections in the next month or so because the price is already beyond their last quarter of 2026 projections and all of the signals are still strong (de-dollarization, geopolitical instability, central bank diversification, dollar inflation, falling interest rates, economic uncertainty). It's a perfect storm that sucks overall but it's good for precious metals.
-
Comment on Microsoft gave US FBI keys to unlock encrypted data in ~society
post_below Link ParentI just used (part of) the title from the article. I'm not attached to it if someone wants to change it. However it's technically accurate, the article talks about specific examples.I just used (part of) the title from the article. I'm not attached to it if someone wants to change it.
However it's technically accurate, the article talks about specific examples.
-
Comment on Microsoft gave US FBI keys to unlock encrypted data in ~society
post_below LinkJust a heads up. For the moment it's still possible to use Windows without being logged in to a MS account and, even if you are logged in, you can choose not to store your bitlocker keys in the...Microsoft confirmed to Forbes that it does provide BitLocker recovery keys if it receives a valid legal order. “While key recovery offers convenience, it also carries a risk of unwanted access, so Microsoft believes customers are in the best position to decide... how to manage their keys,” said Microsoft spokesperson Charles Chamberlayne.
He said the company receives around 20 requests for BitLocker keys per year and in many cases, the user has not stored their key in the cloud making it impossible for Microsoft to assist.
Just a heads up. For the moment it's still possible to use Windows without being logged in to a MS account and, even if you are logged in, you can choose not to store your bitlocker keys in the account.
-
Microsoft gave US FBI keys to unlock encrypted data
37 votes -
Comment on Wilson Lin on FastRender: a browser built by thousands of parallel agents in ~tech
post_below Link ParentIt's too bad to see him doubling down. I finally got around to watching the video interview, or most of it, and in the CNN website part (the only part that wasn't cherry picked by the Cursor dev,...It's too bad to see him doubling down. I finally got around to watching the video interview, or most of it, and in the CNN website part (the only part that wasn't cherry picked by the Cursor dev, likely with pre-cached elements... Simon (or whoever was controlling the cursor at that point) starts to scroll down and quickly stops when it becomes apparent that there's just blank space below the fold. Simon communicated more about his intent by pretending not to notice that than anything he wrote in his post.
-
Comment on Federal officers kill another citizen in Minneapolis, National Guard activated in ~society
post_below Link ParentI don't really have anything useful to add. It's horrifying. Both this time, the last, and the overall frequency of murders. You'd think they would have walked it back for a least a little while...I don't really have anything useful to add. It's horrifying. Both this time, the last, and the overall frequency of murders. You'd think they would have walked it back for a least a little while after Renee Good.
It's like they're pushing harder, to communicate that yes, they actually can get away with executing people with impunity.
-
Comment on Wilson Lin on FastRender: a browser built by thousands of parallel agents in ~tech
post_below Link ParentAdd to this that when people tried to compile and run it shortly after the release, they couldn't get it to compile, nor could the majority of previous versions compile. When it finally did start...Add to this that when people tried to compile and run it shortly after the release, they couldn't get it to compile, nor could the majority of previous versions compile. When it finally did start working there were some unusual commits shortly before that some speculated were actual humans trying to duct tape it together. Disclaimer: that last bit is entirely a rumor as I didn't look at the code or try to compile it myself.
The reason I didn't look, aside from lack of interest, is that I know what GPT and Claude output under the best of circumstances, and it's not something you can mash together into a working browser from scratch. It's not even close.
But with $80,000 in tokens (their estimate), you can get it to pull together a bunch of libraries to do the real work and end up with a demo that works in the sense that you can get it to kind of display a web page but fails to be actually useful for any practical application. A handful of humans could do better in less time with a bar that low.
Willison posts great stuff, I enjoy his blog, but a puff piece is the wrong angle here. This was a publicity stunt for Cursor, relying on the AI crazed tech media not asking too many questions. Simon is an engineer, he could have told a much better story about what Cursor "achieved".
It is a really interesting proof of concept about agents orchestrating themselves, but what it also proves is that even with a blank check and a server farm agents can't make usable, sophisticated software themselves.
Another missing part of the story: Cursor's user base is increasingly vibe coders. Engineers have been switching to better options in droves for at least the last 6 months, which accelerated with the release of Sonnet 4.5 and then Opus 4.5. And started moving even faster when they scrapped their unlimited auto loss leader. So a "demo" like this appeals directly to their target audience of people who can't read the code, and therefore don't know it's slop.
-
Comment on Why does ssh send 100 packets per keystroke? in ~tech
post_below LinkThese are the sorts of coding assistant experiences it's difficult to articulate without context, and also the sort that seem fairly small on their own, but add up to reframing software...I’ve been thinking about whether LLMs remove parts of the problem-solving process that I enjoy. But I’ve gotta say, debugging this problem using Claude Code was super fun.
I am familiar enough with tcpdump, tshark, and friends to know what they can do. But I don’t use them regularly enough to be fast with them. Being able to tell an agent “here’s a weird pcap - tell me what’s going on” was really lovely. And by watching commands as the agent ran them I was able to keep my mental model of the problem up to date.
These are the sorts of coding assistant experiences it's difficult to articulate without context, and also the sort that seem fairly small on their own, but add up to reframing software engineering as a discipline.
I don't know what the percentage is, but at least many developers started doing it because building things is really fucking fun. That and the remarkable truth that you can build something in the digital world that makes something better in the actual world.
But finding bugs is only fun when you can manage a pretty specific mindset. It's great when you have the time and bandwidth to enjoy the journey. But otherwise the best you can say about it is that finding bugs is reliably satisfying and the degree to which it's satisfying is directly proportionate to how many hours you spent figuring it out. You appreciate the outcome more because it hurts a little, which is great, but different than fun.
However finding bugs with the help of a SOTA coding assistant is really fucking fun. Sometimes. They also introduce new kinds of pain.
In the context of coding, it seems like the headlines focus on whether or not AI can one shot things. And that makes sense, it's what the market wants. You can't really blow the lid off of efficiency (read: replace humans) until you can one shot things. But, for right now anyway, these tools real value proposition is removing some of the high friction parts of the process by being effective assistants. Except that doesn't quite cover it because they also open up avenues that were technically always there, but almost never used because they were high friction. That part is really fucking fun too.
-
Comment on Nova Launcher: An update in ~tech
post_below Link ParentIn Smart Launcher you can assign actions to the home button. I currently have double tap on home open the app drawer, it also supports single tap home actions (kicks in if you're already on the...In Smart Launcher you can assign actions to the home button. I currently have double tap on home open the app drawer, it also supports single tap home actions (kicks in if you're already on the main home screen).
Alternatively you can create a shortcut to the app drawer that functions like any other icon.
-
Comment on Let's talk orchestrated objective reduction! in ~science
post_below Link ParentAt the risk of moving the conversation away from physics... If there wasn't enough beauty in the particle interactions, there's also the DNA and RNA, and if there's not enough there, the rest of...To me, its somewhat the opposite: There is such beauty and wonder in how mine own complex self emerges out of the little stochastic determinism of elementary particles
At the risk of moving the conversation away from physics... If there wasn't enough beauty in the particle interactions, there's also the DNA and RNA, and if there's not enough there, the rest of our biology is awe inspiring in its elegance and complexity. If that's not enough, ecosystems are impossibly beautiful and complex webs of interrelation and coevolution.
And we don't understand the larger part of any of it yet. There is potential discovery in every direction.
Which is to say that we have enough wonder, mystery and purpose available in the things we already understand, and the things we can potentially understand with more study, to satisfy even the most demanding spiritual needs for the foreseeable future.
-
Comment on The assistant axis: situating and stabilizing the character of large language models in ~tech
post_below Link ParentAnthropic is, by far, the most transparent and well documented among the frontier model providers. That's how low the bar is unfortunately. On the bright side their strategy is part of the reason...Anthropic is, by far, the most transparent and well documented among the frontier model providers.
That's how low the bar is unfortunately.
On the bright side their strategy is part of the reason they've been leading and everyone else has been copying them when it comes to agents and the all important coding and business tools markets. So hopefully all of the model providers learn from that. You win over developers by giving them technical details and ability to control their workflow, and you win over the rest of the organization by winning over developers.
I'm not sure about alarms but I know someone with hearing issues and I looked into it recently, there are at least a half dozen companies making live caption glasses for hearing impaired people. I agree, it's a fantastic use case.
A few others: research, in so many fields but especially in genomics. Medicine, not to replace doctors but to improve the process (screening, data management, diagnostic aids), modeling (weather, climate, geology, etc..), gaming (the most popular speculation is MUCH better NPCs). Also world building in general, for games of course but the possibilities for "world models" in a variety of areas are near unlimited. If you can train on the "world" of a domain (say, an ecosystem), the resulting inference could be remarkably useful. I expect there will be a lot of annoying hype around world models this year.
Regarding spiraling, I get it. I don't want to take this thread into the downsides direction but they're as unlimited as the upsides.
In spite of that, it's going to happen no matter what we think about it. We can hopefully support the creation of regulation and guardrails but one way or another the technology is going to keep exploding in all directions.
If some day we look back and collectively decide that AI was a huge mistake, not one of us will be able to realistically say "I should have done more to stop it" because there's just not currently anything an individual can do relative to the unprecedented amount of capital involved. Even large groups of individuals don't stand a chance.
I don't say that from a defeatist perspective. At any point in history there are large scale developments and circumstances beyond an individual's control. With just the one life each, I think we should appreciate the upsides where we can and focus our energy in the places we actually can make an impact. There have always been too many problems for one person to solve.