macNchz

Gemma 4 26B-A4B might be interesting to try on your machine. The latest optimizations make MoE models work pretty nicely on setups like that with a decent GPU and lots of slowish RAM. I have a 16gb GPU and 64gb of 3200mhz DDR4 and get 15-20 tokens/sec out of that model with zero finagling or tweaking. I’ve been very impressed by it, even having run just about every other open weight model that would fit on my machine over the last few years.

If you're building on top of APIs and can do some eval work (aka do not need the most bleeding edge model), the Gemini Flash and Flash Lite models are super capable for the price.

This is something I've been thinking about for a while...the current state of things really does feel kind of like the dialup era, wondering what the "broadband" era could look like. Watching tokens stream in is reminiscent of watching a jpeg load a few rows of pixels at a time, and the various different loading and connecting animations that applications implemented before things got fast enough to make them less relevant.

Some of the work in that direction like Cerebras or Taalas have been doing is an interesting glimpse of what might be possible. In the meantime it's a fun thought experiment to wonder about what might be possible if even current state of the art models were available at like, a million tokens per second at a very low cost.

They were arbing keywords through this: they'd place ads directly on Google for low-cost keywords, which would link to an ask.com search results page that itself would display Google ads through that partnership, but with a UI designed (more than Google itself) to trick people into clicking them. Seemingly they were able to find combinations that made this profitable.

The "Search Partner network" in general is one of the ways that Google Ads milks (scams?) unsophisticated buyers: unless you turn it off, you're paying for ads that are shown to confused users on sketchy results pages that you have no real insight into, not just the Google results page itself. The traffic from them is garbage.

There are plenty of ways you could wind up with a git commit containing "OpenClaw" despite zero interaction with OpenClaw itself...adding a blog post to a static site repo, or even a clause in your own app's ToS disallowing use of OpenClaw with your API.

Yeah I actually find writing the prompt itself to be such a useful mechanism of thinking through problems that I will not-infrequently find myself a couple of paragraphs in and decide to just delete everything I've written and take a new tack. Only when you're truly outsourcing your thinking to the AI will you run into the situation that the LLM being down means you can't actually work at all.

An interesting element here, I think, is that writing has always been a good way to force you to organize and confront your thoughts. I've liked working on writing-heavy projects, but often in fast-moving environments writing things out before coding becomes easy to skip over, but working with LLMs has sort of inverted that. You have to write to produce code with AI (usually, at least), and the more clarity of thought you put into the writing the better the outcomes (usually).

I recently broke my 12 Mini beyond reasonable repair during a battery replacement recently (mostly just bad luck, I've been doing my own mobile repair for a long time). I bought a 17 to replace it, and promptly returned it in favor of a used 13 Mini. It's wild to me how large the smallest mainstream phone you can buy these days has become.

This is something a lot of people seem to believe that is not borne out in the research. Plenty of specific counter examples like Queens NY, a densely populated and exceptionally diverse place with crime rates comparable or better than many much more homogenous places in America. Poverty and income inequality are much better predictors. I felt this reddit comment from a while ago did a pretty good job rolling up sources on this: https://www.reddit.com/r/TrueUnpopularOpinion/comments/1jxff...

Yes absolutely—I think cars have obvious utility as machines, but there has now been 100 years of building everything around them and changing laws in such a way that encourages their use: through direct and indirect subsidy, land use rules that largely outlaw building cities in any way other than sprawl that itself increases the importance and utility of cars, and various other preferential regulations that often tolerate the harms in a way that is not applied elsewhere (c.f. panic over e-bike safety vs American highway safety overall).

Yeah there's a weird thing where people would get really focused on whether something is "actually doing RAG" when it's pulling in all sorts of outside information, just not using some kind of purpose built RAG tooling or embeddings.

Now, the pendulum on that general concept seems to be swinging the opposite direction where a lot of those people just figured out that you don't need embeddings. That's true, but I'd suggest that people don't overindex on thinking that means embeddings are not actually useful or valuable. Embeddings can be downright magical in what you can build with them, they're just one more tool at your disposal.

You can mix and match these things, too! Indexing your documents into semantically nested folders for agents to peruse? Try chunking and/or summarizing each one, and putting the vectors in sidecar files, or even Yaml frontmatter. Disks are fast these days, you can rip through a lot of files indexed like that before you come close to needing something more sophisticated.

I’ve created a bunch of fresh Azure accounts over the past few years and each time I’ve found myself sitting there dumbfounded anew at how garbage the experience is.

There has been weird broken jank at just about every step of the process at one point or another. Like, I’m a serious person trying to set something up for a production workload, and multiple times along the way to just having a working account that I can log into with billing configured, I’ll get baffling error messages like [ServiceKeyDepartureException: Insufficient validation expectancy. Sfhtjitgfxswinbvgtt-33-664322888], and the whole thing will simply not work until several hours later. Who knows why!?

I evaluated some Azure + Copilot Studio functionality for a project recently, which required more engagement with their whole 365 ecosystem than I’d had in a long time and it had many of the same problems but worse. Just unbelievably low quality software for the price and how popular it is. Every step of the way I hit some stupid issue. The people using this stuff are clearly not the people buying it.

Beyond just invasive/annoying, ad networks explicitly spread malware and scams/fraud. There's not much incentive for them to clamp down on it, though, as that would cost them money both in lost revenue and in paying for more thorough review.

It’s kind of baffling to me that laptops in classrooms took off the way they did, as it seemed like a distraction machine to me even 25+ years ago, as a kid myself! My school got some carts of laptops that would move from classroom to classroom in ~2000—they were heavily used for flash games and other nonsense, and were strictly worse for that than in the dedicated computer lab classroom, where all of the monitors faced into the center of the room where the teacher could see them.

When I got to college a few years later I’d sit in the back of classrooms and see that a majority of students who’d brought a laptop (ostensibly for notes) were consistently distracted and doing something else, be it games or StumbleUpon. I can only imagine these decisions were made by groups of adults sitting around conference rooms, each staring at their own laptop and paying 20% attention to the meeting at hand.

I find that all of Google’s ad products are under-moderated for malicious ads. It’s a choice on their part to not tightly control this—they certainly could, though it would harm their incredible profitability if they did more scrutiny on the ads they show. I personally don’t especially care to pay a premium not to see deepfakes of celebrities promoting crypto scams.

There's definitely a productivity curve element to getting it to behave effectively within a given codebase. Certainly in the codebases I work with most frequently I find Claude will forget certain key aspects (how to run the tests or something) after a while and need a reminder, otherwise it gets into a loop like that trying to figure out how to do it from first principles with slightly incorrect commands.

I think a lot of the noise about letting Claude run for very extended periods involves relatively greenfield projects where the AI is going to be using tools and patterns and choices that are heavily represented in training data (unless you tell it not to), which I think are more likely to result in a codebase that lends itself to ongoing AI work. People also just exaggerate and talk about the one time doing that actually worked vs the 37 times Claude required more handholding.

The bigger problem I see with the "leave it running for the weekend" type work is that, even if it doesn't get caught up on something trivial like tabs vs spaces (glad we're keeping that one alive in the AI era, lol), it will accumulate bad decisions about project structure/architecture/design that become really annoying to untie, and that amount to a flavor of technical debt that makes it harder for agents themselves to continue to make forward progress. Lots of insidious little things: creating giant files that eventually create context problems, duplicating important methods willy nilly and modifying them independently so their implementations drift apart, writing tests that are..."designed to pass" in a way that creates a false sense of confidence when they're passing, and "forest for the trees" kind of issues where the AI gets the logic right inside a crucial method so it looks good at a glance, but it misses some kind of bigger picture flaw in the way the rest of the code actually uses that method.

I've been wondering if we'd see a cyber campaign emerge in this conflict. To my knowledge Iran seems to have pretty advanced cyber capabilities and increasingly fewer reasons to hold back. Gloves-off cyber war doesn't sound good to me. The US CISA already been cut back, has lost "virtually all of its top officials"^, doesn't have a permanent director, and is operating at a further reduced capacity because of the DHS shutdown.

^ https://www.cybersecuritydive.com/news/cisa-senior-official-...

Opus 4.5 to 4.6 was pretty incremental, I didn't see much of a difference.

The big coding model moments in recent recollection, IMO, were something like:

- Sonnet 3.5 update in October 2024: ability to generate actually-working code using context from a codebase became genuinely feasible.

- Claude 4 release in May 2025: big tool calling improvements meant that agentic editors like Claude Code could operate on a noticeably longer leash without falling apart.

- Gemini 3 Pro, Claude 4.5, GPT 5.2 in Nov/Dec 2025: with some caveats these were a pretty major jump in the difficulty and scale of tasks that coding assistants are able to handle, working on much more complex projects over longer time scales without supervision, and testing their own work effectively.

Watching over shoulders as elderly people watch YouTube with ads and engage with clips of deepfake celebrities selling fraudulent nonsense is both enlightening and painful.

Not claiming to have hit on something unique here, but I think it’s realistic and often drowned out in favor of sci-fi nonsense.

It's definitely creeping into things, though most of the features I've seen are fairly simplistic compared to what would be possible if the video was being reviewed + indexed by current SoTA multimodal LLMs.

Presumably the ideal is some kind of a fusion. Upload or tag some images/videos and link someone's social profiles and the system can look out for them based on facial recognition, gait recognition, vehicle/pets/common wardrobe items in combination.

This is a really cool implementation—embeddings still often feel like magic to me. That said, this exact use case is sort of also my biggest point of concern with where AI takes us, much more so than most of the common AI risks you hear lots of chatter about. We live in a world absolutely loaded with cameras now but ultimately retain some semblance of semi-anonymity/privacy in public by virtue of the fact that nobody can actually watch or review all of the video from those cameras except when there is a compelling reason to do so, but these technologies are making that a much more realistic proposition.

The presence of cameras everywhere is considerably more concerning than the status quo, to me at least, when there is an AI watching and indexing every second of every feed—where camera owners or manufacturers or governments could set simple natural language parameters for highly specific people or activities notify about. There are obviously compelling and easy-to-sell cases here that will surely drive adoption as it becomes cost effective: get an alert to crime in progress, get an alert when a neighbor who doesn't clean up after his dog, get an alert when someone has fallen...but the potential implications of living in a panopticon like this if not well regulated are pretty ugly.

Was curious—good number of projects out there with an un-pinned LiteLLM dependencies in their requirements.txt (628 matches): https://github.com/search?q=path%3A*%2Frequirements.txt%20%2...

or pyproject.toml (not possible to filter based on absence of a uv.lock, but at a glance it's missing from many of these): https://github.com/search?q=path%3A*%2Fpyproject.toml+%22%5C...

or setup.py: https://github.com/search?q=path%3A*%2Fsetup.py+%22%5C%22lit...

I dunno, I thought about this before switching to Linux, when I gave my wife a Linux box I had sitting around in a pinch during the pandemic laptop shortage—a lot of people these days just need a browser, and there’s not really much to go wrong with that. If something does go wrong you can just nuke the whole thing and start over pretty easily.

I’ve certainly run into some odd situations on my desktop Linux machine over the past 6 years since I started using it full time, but I think most of them were related to the nature of how I use the machine more than inherent instability. I think I’ve spent many more hours of my life unwinding piles of malware and bloat from non-technical folks’ Windows machines than debugging this one.

Great point. Just last week I used AI to build a minimal replacement for a SaaS tool I’ve used in the past that has obnoxious feature gating/price tiers. My version isn’t nearly a complete replica, but it has the base functionality I want without having to feel like someone spent hundreds of hours perfecting price tiers with artificial limitations that annoy me just enough to upgrade.

Getting a tool that did exactly what I wanted with no fuss was delightful.

The only places I've heard of losing water during power outages are houses that use a private well (no power, no well pump), which would be the case anywhere. Municipal water systems may or may not use power to provide pressure, but are going to have generator power outside of the most severe outages.

I use uv here and there but have a bunch of projects using regular pip with pip-tools to do a requirements.in -> requirements.txt as a lockfile workflow that I've never seen enough value in converting over. uv is clearly much faster but that's a pretty minor consideration unless I were for some reason changing project dependencies all day long.

Perhaps it never grabbed me as much because I've been running basically everything in Docker for years now, which takes care of Python versioning issues and caches the dependency install steps, so they only take a long time if they've changed. I also like containers for all of the other project setup and environment scaffolding stuff they roll up, e.g. having a consistently working GDAL environment available instantly for a project I haven't worked on in a long time.

There’s a name for that, SLAPP: https://en.wikipedia.org/wiki/Strategic_lawsuit_against_publ...

Many states in the US have laws to try to limit them by making them easier to dismiss etc.

No maliciousness or injection required, even the newest and most resistant models can start doing weird stuff on their own, particularly when they encounter something failing that they want to work.

Just today I had Opus 4.6 in Claude Code run into a login screen while building and testing a web app via Playwright MCP. When the login popped up (in a self-contained Chromium instance) I tried to just log in myself with my local dev creds so Claude would have access, but they didn't work. When I flipped back to the terminal, it turned out Claude had run code to query superadmin users in the database, picked the first one, and changed the password to `password123` so it could log in on its own.

This was a sandboxed local dev environment, so it was not a big deal (and the only reason I was letting it run code like that without approval), but it was a good reminder to be careful with these things.

In legal states in the US it’s fairly easy to find commercial 1:1 THC:CBD products, among a litany of mega-potent THC-only options.

I never particularly cared for cannabis in general in the past–it didn’t give me panic attacks, but even a little bit made me feel kind of on edge. The 1:1 stuff is dramatically more pleasant.