Deno Desktop looks quite attractive:
- Use OS WebView with a JavaScript backend - a lower barrier to entry than Tauri or bundle with embedded Chromium
- Super-simple setup and Deno’s rich tooling
One great early use case could be building custom dev tooling apps with AI agents
This is only compute time though, moving data between CPU and GPU takes another 41ms which brings it close to compute time on CPU. Interesting to compare on tasks that require less communication and more compute and try devices with unified memory (like Mac M-series)
Quite fascinating to see how GPUs can be crazy fast on certain tasks. Summing up two vectors (arrays) each consisting of 16.7M elements is taking less than 1ms, while on CPU it takes 54ms (4 vCPU)
🚨 I'm really excited to share a super early look at my new project.
As you may have noticed, I've been writing some stuff around quantum computing recently. Instead of continuing this in my personal blog, I've decided to start collecting these articles on a new site.
QCFundamentals will begin as just a collection of these articles but eventually expand into a more complete course. I also want to add some fun interactive tools in the future.
🔗Check it out!
https://t.co/sARKEqc6df
Modelling schemes from "The Manga Guide to Microprocessors" with logisim-evolution app. It's quite nice for simulating combinations of inputs and extracting part of scheme to a reusable module.
For example schemes for half adder and full adder for adding binary digits:
The second order effect of the 2nd point is that the web will no longer be beholden to HTML (for real this time bc SEO and a11y will change)
At a time where folks double down on some of the newest JS framework techniques, I actually think they’ll go obsolete
Nice post by @mattrickard about demand shifting towards full-stack engineers and specialisation outsourced to third-party SaaS. Although front-end have been particularly resistent to adoption of various web app constructors (web-platform changing too fast) https://t.co/88uvteS2T5
We’ll come full hilarious circle when people use LLMs both to 1) expand a simple message like “execute faster” into email and 2) summarize an email back into the original simple message. It’s like compression/decompression into formalese
Not sure if there is a name for (I think no) the feeling of a deep discomfort when the probability of an interruption is > 0 while trying to work. It’s a kind of fear.
На 10 апреля ООН подтвердила 142 убитых и 229 раненых детей.
"Одного из таких детей я встретил в реанимации в больнице в Запорожье. Четырехлетний Влад получил два выстрела в живот, когда убегал с семьей от обстрелов. Ожидается, что Влад будет жить, в отличие от многих других"