crawshaw

8 ms·

91.
by crawshaw 10mo ago
We are risking word games over what can make competent decisions, but when my thermostat turns on the heat I would say it decided to do so, so I agree with you. If someone has a different meaning of the word "decided" however, I w
92.
by crawshaw 10mo ago
I do not have a useful opinion on another person’s emotional response. My post you are responding to is about responsibility. A legal entity is always responsible for a machine.
93.
by crawshaw 10mo ago
I certainly have no intention of doing anyone harm. I went to their website and clicked three times to get the names of the people and organization behind it, there is a prominent About page with profile links. If an admin considers this in
94.
by crawshaw 10mo ago
The important point that Simon makes in careful detail is: an "AI" did not send this email. The three people behind the Sage AI project used a tool to email him. According to their website this email was sent by Adam Binksmith, Za
95.
by crawshaw 10mo ago
Data races are a source of bugs. They are not a noticeable fraction of the security issues that face memory unsafe languages, which is the practical argument for memory safety.
96.
by crawshaw 10mo ago
I am not sure if you are: 1. attempting to retcon garbage collected languages as not memory safe, or 2. discussing a particular implementation choice of the standard Go runtime that was made because it is not a practical source of bugs (see
97.
by crawshaw 10mo ago
The funny thing is if GitHub let me pay extra for an actions runner that was not a potato, I would happily give them so much money. Instead they want to penalize me for working around their broken product.
98.
by crawshaw 10mo ago
This does not appear to be true. Six months ago I created a small programming language. I had LLMs write hundreds of small programs in the language, using the parser, interpreter, and my spec as a guide for the language. The vast majority o
99.
by crawshaw 11mo ago
That is clearly not true. SLS is much more expensive by any measure and is not reusable in any way. Other interesting work, e.g. rocket lab, is not old space.
100.
by crawshaw 11mo ago
Has anyone who is a regular Opus / GPT5-Codex-High / GPT5 Pro user given this model a workout? Each Google release is accompanied by a lot of devrel marketing that sounds impressive but whenever I put the hours into eval myself it
101.
by crawshaw 11mo ago
This long predates AI codegen and I believe discussing it here is a distraction. Projects written to grab mindshare rather than to be directly used by those building them have been around for years. These projects exist to get attention. I
102.
by crawshaw 11mo ago
exe.dev | Backend and Frontend | SF Bay Area | Full-time Doing fun new things with VMs. Details: https://exe.dev/jobs
103.
In Defense of the Amyloid Hypothesis(astralcodexten.com)
2 points by crawshaw 1y ago|0 comments
104.
by crawshaw 1y ago
It is a “trap” in the sense that it limits the design space for your language by forcing some of the choices implicit in C++. Rust went with slow build times like C++ and so was not held back in this dimension by LLVM. Nothing about LLVM is
105.
by crawshaw 1y ago
exe.dev | Product Engineer | Hybrid ~2x/wk | SF | Full-Time Building compute. $150k-$200k. Early stage decent equity.
106.
by crawshaw 1y ago
The most interesting point in this post, which resonates with me, is to those of us who work a lot, 996 sounds ridiculous. It sounds ridiculous because to work a lot, you have to fit in the gaps around your life. I have done about 60hrs&#x
107.
by crawshaw 1y ago
It is great that Fil-C exists. This is the sort of technique that is very effective for real programs, but that developers are convinced does not work. Existence proofs cut through long circular arguments.
108.
by crawshaw 1y ago
> All the innovation is coming "top-down" from very well funded companies - many of them tech incumbents The model leaders here are OpenAI and Anthropic, two new companies. In the programming space, the next leaders are Qwen an
109.
by crawshaw 1y ago
A title of "Memory is getting slower, Disk is getting faster" might remove a lot of the initial confusion this post causes.
110.
by crawshaw 1y ago
sketch.dev | engineer | onsite (SF) two days a week | Full-Time | $200,000 + generous equity We are expanding the hosted side of our product and are looking for people who enjoy "cloud" infrastructure, have a strong product instin
111.
by crawshaw 1y ago
A bubble is asset prices systematically diverging from reasonable expectations of future cash flows. Bubbles are driven by financial speculation. The claim in the blog post that all technology leads to speculative asset bubbles I find hard
112.
by crawshaw 1y ago
If you believe everything around you must be good before you can do good, then you must trick yourself into believing imperfect things are good to do anything, or withdraw completely from the world. Your radical philosophy leads either to d
113.
by crawshaw 1y ago
No, it is about doing what you can, with what you have, where you are. That is good.
114.
by crawshaw 1y ago
Nice example of how good programming is often about meeting systems where they are: Whatever your feelings are about OOP and memory management, the reality is that if you choose GTK, you're forced into interfacing in some way with
115.
by crawshaw 1y ago
One of the sketch.dev folks here, if you have any questions. I know containers are hard to use today, and I don't think we have found the right magic yet. But we're working on it, because all of these tools are so much better in a
116.
by crawshaw 1y ago
If you would like to try this in a coding agent (we find the qwen3-coder model works really well in agents!), we have been experimenting with Cerebras Code in Sketch. We just pushed support, so you can run it with the latest version, 0.0.33
117.
by crawshaw 1y ago
If you want to see it in action in some code, our agent Sketch uses a TODO list tool: https://github.com/boldsoftware/sketch/blob/main/claudetool/... It is relatively easy to get the agent to use it
118.
by crawshaw 1y ago
This is a great blog post. I have seen internal studies at software companies that demonstrate this, i.e. reducing UI latency encourages more software use by users. (Though a quick search suggests none are published.)
119.
by crawshaw 1y ago
I first got paid to code 25 years ago. I have been programming since I was a kid, and always will. I love making computers do things. I definitely don't love the process: design docs, meetings, code review, CI, e2e tests working around
120.
by crawshaw 1y ago
All bets aren’t off, we empirically measure the safety of software based on exploits. C memory handling is most of its exploits. Show me the exploits based on Go parallelism. This issue has been discussed publicly for 10 years yet the explo

More ›