doug_durham

4 ms·

61.
by doug_durham 4mo ago
There’s nothing new here. This is how open source software has been since its inception. It’s just the nature of reality.
62.
by doug_durham 4mo ago
As has been pointed out elsewhere, DMA isn't a privacy regulation. It is simply about competition. You can be in 100% compliance with DMA and poor privacy protections. This is the crux of the problem. How do you preserve the privac
63.
by doug_durham 4mo ago
I have to lift up my hand and move my arm around to use it. With a trackpad all I need to do is move my hand over and flick my fingers for gestures. My wrist never moves.
64.
by doug_durham 4mo ago
Rust, Python, Java, Ruby, Scala, Swift to start with. These are languages with very wide adoption. Objective-C is very Smalltalk-like, but it is being phased out for Swift.
65.
by doug_durham 4mo ago
This is a truly important paper. It formalizes the intuition that many in the field have. We can stop wasting time doing formal analysis of LLMs. If you have a problem that requires formal verification, don't use an LLM. You can use
66.
by doug_durham 4mo ago
You think that LISP and Smalltalk aren't widely used is because they weren't easy to implement in the late 1980's? There have been many languages that have risen to prominence in the 40 years since, yet LISP and Smalltalk re
67.
by doug_durham 4mo ago
I use a trackpad to avoid virtually all of the issues created by a mouse. The trackpad gestures in macOS are magical.
68.
by doug_durham 4mo ago
Work has never been about "discovering the world". There have been a handful of privileged folks who had the time to "discover the world". Work has traditionally been "let's find enough food for my family&quo
69.
by doug_durham 4mo ago
DSLs. Creating a language that only you know that will double the learning curve for the folks coming after you. It's fine for personal projects, but almost always an anti-pattern.
70.
by doug_durham 5mo ago
There is nothing original in this article. This could’ve been written by an AI just scraping one month of posts to hacker news. This article is a critique of autonomously repackaging existing ideas by autonomously repackaging existing idea
71.
by doug_durham 5mo ago
It really depends on your goal. If your goal is to spend the evening coming up with funny things to say with your friends, then you shouldn't use AI. If your goal is to finish the t-shirts so that you can move on to the next topic in
72.
by doug_durham 5mo ago
I find these kinds of posts to be elitist and self-important. They draw a false dichotomy between tool use and lifestyle. I'm glad the poster has a lifestyle that works for them personally. This post really has nothing to do with AI.
73.
by doug_durham 5mo ago
Why is it a sacred journey? They are quitting a job at Sentry and taking one a Home Depot. As much as I value the role that Home Depot plays in society I'd never use the word "sacred" to describe the work, nor the work at a
74.
by doug_durham 5mo ago
GLM-5.1 isn't just as good. It is no match for Opus running in Claude Code. Please try it yourself. Open source models are about a year behind at least.
75.
by doug_durham 5mo ago
Open source models that you can run locally are much more than 3 to 6 months behind. 6 months was the November inflection for Claude. No open source model is as good as Claude Opus 4.6.
76.
by doug_durham 5mo ago
The space of palatable human food is small. There are only a few thousand ingredients and a few thousand preparation techniques. This could easily be compressed at high fidelity into a model.
77.
by doug_durham 5mo ago
I'm sorry what do you expect? Did you expect people to not use Google search prior to AI tools? People don't know the answer to your question and they are trying to be helpful. Not everyone is effective at using AI tools at thi
78.
by doug_durham 5mo ago
Yes but... No one living today knows what direction AI technology will take humanity. If we have an algorithm breakthrough then we may avoid building new data centers. If the abilities of the technology plateau then there might not be lar
79.
by doug_durham 5mo ago
What do you find yourself gravitating to? What part of your job comes easiest? That things are easy to you that other’s find difficult? What do you spend time learning more about even when you don’t have to? Those are directional. For
80.
by doug_durham 5mo ago
So transcoding doesn't work unless every line of code is read? That's not how transcoding is done in practice.
81.
by doug_durham 5mo ago
Really?? So you base your engineer in "speculation". The Bun team has a deep track record of delivering a high quality product. What makes you think that is going to stop?
82.
by doug_durham 5mo ago
I'm certain that the maintainers of Bun have excellent understanding of their codebase. What makes you think that they don't? They wrote the code in the first place. They know the architecture. They know what pieces do what fu
83.
by doug_durham 5mo ago
There is no evidence that it was "vibe" coded. It was ported to Rust by an expert engineer using an AI tool using solid SWE practices.
84.
by doug_durham 5mo ago
Yes, this is the standard model for the big frontier models. You don't need Gemini or Claude to do tab completions. A modest size local model can do that just fine. If that is all you are using AI tools for you are wasting money sub
85.
by doug_durham 5mo ago
You have to wait in line behind several people to get to the point where you can talk to the cashier. Most McDonalds have several kiosks. There is usually little or no line. I can place my order grab a table.
86.
by doug_durham 5mo ago
I write low level highly performance sensitive code and I get excellent results using LLMs. It isn't a narrow area of applicability.
87.
by doug_durham 5mo ago
There seems to be a split. There are a substantial group of people like yourself who don't get what they want from LLMs. There are folks like me who get excellent idiomatic code almost every time. Both experiences are real and shoul
88.
by doug_durham 5mo ago
You are choosing to be enraged. If a tool works it works.
89.
by doug_durham 5mo ago
Accountability is the key. You are accountable for the work product. You need to do what is necessary to make sure that the product meets standards. How you do it is of secondary importance. If reading code is what is needed to guarantee
90.
by doug_durham 5mo ago
That presumes that software has a wide solution space. In my experience writing software from the kernel all of the way to distributed systems is that the space is narrow. There are certainly a few long tail areas that require unique appr

More ›