Tor

Working on some pull-ups. Getting older, but still have some gas left in the tank! #pullups

@ScriptedAlchemy I am running qwen 3 0.6B locally. Free and nice 😊

If you’re using local models, you’re not solving complex enough problems.

@chrisalbon They serve a good Philly cheese steak

Okay I’ve never been to a Jersey Mike’s. What am I missing here?

One of my recent side projects was building a personal chatbot that can answer questions about my household. As part of that, I experimented with fine-tuning a local LLM to categorize incoming questions to provide better RAG query results teachmecoolstuff.com/viewarticle/fi…

@bernhardsson @valeriibo If it can’t be tied to meaningful productivity gains, it’s definitely a bubble

@valeriibo Yes correct

I think the AI layoffs are real but in a weird way – it’s about reallocating resources to AI. Either within the company (Meta, Microsoft), or because capital markets are rotating out of traditional software (Duolingo, Klarna, etc) towards “AI stocks”

@StatisticsFTW Thanks! Looks like vite is very popular these days.

@HelgevoldTech pnpm + turborepo + oxfmt + oxlint + vitest + vite is a good set of defaults. vite vs nextjs vs whatever is a decision that's harder to undo, so that is worth researching your options and seeing what appeals to you.

What is the most sensible tool chain if you are starting a React application today? #react #frontend #javascript

@kenwheeler What happened?

i truly and honestly wish tech would decentralize from this hive of weird

i’d literally rather die in a coal mine collapse than participate in sf tech culture. it is the cringiest shit, i couldn’t even make it up.

@simonw I suspect we will transition into a period of token budgeting regardless… Similar to how cloud spend transitioned, but on a larger scale..

Here's an extended edit of the quote that includes a following fragment where Andrew Macdonald called the trade "harder to justify" - full, unedited transcript is here: gist.github.com/simonw/59096a3…

I'm suspicious of that that whole story about Uber blowing their AI budget and being disappointed in the results - I dug into it and it appears to have been built on very shaky foundations

@MancerAI_ @davepl1968 @RealMikeWeed The original tweet is in reference to everyone who uses MS word, chrome etc as their primary function. Paying a premium is completely wasted on everyday folks…

@davepl1968 @RealMikeWeed Fully agree. The build quality, track pad and now lately ofc the M-SoC are great. Between OS Cw, windows and Linux I don't have super strong opinions but I'd put OSX last

There are many bad "hot takes", but "the MacBook isn't worth it" is incredibly naive. Or to put it in terms he'd understand, there's an actual difference between Seville Row and buying off the rack.

@AdamRackis This is a bit like when companies spent like drunk sailors on cloud apps.. Then initiated cost cutting measures after receiving the first few bills

Well this has me horny af

@davepl1968 Is humidity a problem for the computers in a garage?

I'm pleased to announce that, when combined with my phone, my PDP-11 has achieved a storage capacity of 128.01GB!

@unclebobmartin We are still looking at the poor man's version of the capabilities portrayed in Star Trek.

These agents are the Star Trek computers. You can ask them question and get answers just like Spock and Kirk used to do. It's uncanny.

@InsiderPresider @jobergum In my experience vector searches struggle with chat questions involving fictional characters and general named entities. Other techniques can be used as a fallback .

@jobergum where

As your bm25 guy its on brand to tell you all that i'm going to ai engineer world's fair to talk about bm25

@Avicula11 @jobergum Isn’t full fine tuning too resource intensive to be practical for most purposes? More lightweight fine tuning options don’t seem like appropriate replacement for RAG based options though

@jobergum both have there own use cases. Finetuning when you want data to get merged with model itself. RAG when we want to use some dynamic information and pass it to llm dynamically.

Finetuning versus RAG debate

@jfversluis @gholdridge Many are likely still running .net framework 4.8…

@gholdridge Love it!

What's the oldest .NET version still running in your production? No judgment. Just curious how far back some of us are holding the line.

There will be no AI jobpocalypse. The story that AI will lead to massive unemployment is stoking unnecessary fear. AI — like any other technology — does affect jobs, but telling overblown stories of large-scale unemployment is irresponsible and damaging. Let’s put a stop to it. I’ve expressed skepticism about the jobpocalypse in previous posts. I’m glad to see that the popular press is now pushing back on this narrative. The image below features some recent headlines. Software engineering is the sector most affected by AI tools, as coding agents race ahead. Yet hiring of software engineers remains strong! So while there are examples of AI taking away jobs, the trends strongly suggest the net job creation is vastly greater than the job destruction — just like earlier waves of technology. Further, despite all the exciting progress in AI, the U.S. unemployment rate remains a healthy 4.3%. Why is the AI jobpocalypse narrative so popular? For one thing, frontier AI labs have a strong incentive to tell stories that make AI technology sound more powerful. At their most extreme, they promote science-fiction scenarios of AI “taking over” and causing human extinction. If a technology can replace many employees, surely that technology must be very valuable! Also, a lot of SaaS software companies charge around $100-$1000 per user/year. But if an AI company can replace an employee who makes $100,000 — or make them 50% more productive — then charging even $10,000 starts to look reasonable. By anchoring not to typical SaaS prices but to salaries of employees, AI companies can charge a lot more. Additionally, businesses have a strong incentive to talk about layoffs as if they were caused by AI. After all, talking about how they’re using AI to be far more productive with fewer staff makes them look smart. This is a better message than admitting they overhired during the pandemic when capital was abundant due to low interest rates and a massive government financial stimulus. To be clear, I recognize that AI is causing a lot of people’s work to change. This is hard. This is stressful. (And to some, it can be fun.) I empathize with everyone affected. At the same time, this is very different from predicting a collapse of the job market. Societies are capable of telling themselves stories for years that have little basis in reality and lead to poor society-wide decision making. For example, fears over nuclear plant safety led to under-investment in nuclear power. Fears of the “population bomb” in the 1960s led countries to implement harsh policies to reduce their populations. And worries about dietary fat led governments to promote unhealthy high-sugar diets for decades. Now that mainstream media is openly skeptical about the jobpocalypse, I hope these stories will start to lose their teeth (much like fears of AI-driven human extinction have). Contrary to the predictions of an AI jobpocalypse, I predict the opposite: There will be an AI jobapalooza! AI will lead to a lot more good AI engineering jobs, and I’m also optimistic about the future of the overall job market. What AI engineers do will be different from traditional software engineering, and many of these jobs will be in businesses other than traditional large employers of developers. In non-AI roles, too, the skills needed will change because of AI. That makes this a good time to encourage more people to become proficient in AI, and make sure they’re ready for the different but plentiful jobs of the future! [Original text in The Batch newsletter.]

@ThePrimeagen What's the latest issue?

My continual disgust at the JS ecosystem feels vindicated practically daily at this point

@complex_maths @AdamRackis I had never heard of it despite bun using it…

@AdamRackis Zig existed before Bun. Ghostty is written in Zig. I think it will still continue to grow for those who want to use ultra-low level, ultra-performant code (especially if coded by hand), but rust will still be the preferred choice when used by AI due to enforced memory safety.

Now that Bun is moving off Zig ... will we ever hear about that language again? Bun was the only thing I'd ever heard of using it, and it's moving to Rust. Hard to imagine anyone being eager to build on Zig at this point. Am I missing something?

@RyanCarniato A fully working React app is the new console log … That’s how cheap it is now to generate a UI as part of random POCs or spikes these days…

The best part of AI is how much it separates hard (worthwhile) design from tedious work. Changes the value of things. Roughly equal amount of code: Perfect Async Reactive System ~ 1 Year Solid's JSX Compiler in Rust ~ 1 Day

@HeadmasterDuck @AdamRackis You can’t really avoid giving it coding instructions alongside requirements. Otherwise you end up with your whole app in a single 5000 line python file… I have seen it happen 😂

@AdamRackis Agreed, for the record. Do you feel like the activities operate at different cognitive levels though? I feel like I can express a detailed spec in spoke ln form with less cognitive overhead than the act of coding which requires that intense flow state for me.

Well, this, yes. And also … depending on what exactly you need to accomplish a sufficiently detailed specification may be more time and effort to write (and wait on) than just changing the code yourself.