kohlerm

Yeah well. MCPs are better for use cases where remote access is required, but for development use cases what you need in the majority of use cases is to manipulate local files. Skills are just the more natural solution here. You can argue whether Skills should come with more type information (MCPs are slightly better here), but otherwise it seems pretty clear to me that if you do not need remote access then MCPs are not really needed.

I checked the codex source code a few months ago and the implementation was very basic compared to opencode

It is definitely a real problem to be solved. We used something similar in an IOT application some time ago

Very good article. With regards to the guy being a designer, he is IMHO still correct with regards to layouts. Currently LLMs ares till pretty clueless about layouts. Also SWE is much more then coding. Even in the coding area there is much more room for improvement. The idea that you would not need Software Engineers soon, is brain dead.

I think I have seen it being used in Snowboards

All politicians follow their beliefs No they don't they, a lot them will do whatever is needed (change their minds at will) to stay in power. If you ever listened to Habeck, there are simple rational arguments why it did not makes sense to continue running the nuclear plants for longer. There was anyway no easy affordable way to just let them run for many years (contracts). At the end nuclear was only contributing 1.5% to the overall energy production.

Merkel (CDU) pushed it after Fukushima, because "oh it's dangerous", which wasn't rational, but she knew people wanted it

He is on of the few politicians who follow his beliefs about what needs to be done, less so about what is popular. He also made sure Germany had enough gas. He had nothing todo with the pushback of nuclear energy (that was the CDU/Merkel). Nuclear energy is not cheap.

Cursor's search is still better(faster and cheaper) then Claude Code. I just did some tests. It looks like they do agentic searches with query rewriting.

Following dependencies is the way to go IMHO. Saying "Code Doesn't Think in Chunks", is IMHO not correct. Developers do thing in chunks of codes. E.g. this function calls that function uses that type and is used here and there. It is not really a file based model like Cline uses. The file based model is "just" simpler to implement :-) . We use a more sophisticated code chunking approach in https://help.sap.com/docs/build_code/d0d8f5bfc3d640478854e6f... Let's see, maybe we should open source that ...

I am not so sure this makes sense. Training a model to directly use certain tools (web search etc) makes the model very specialized and less flexible. As long as other solutions are more flexible and less costly, training a specialized model would not be worth it.

it is also the "right thing to do" IMHO.

I am using a slightly modified Extend layout since a couple of years, and it was really a game changer for me.

isn't 1080p resolution too low for that size?

Nice! would have needed something like this last year.

I agree that putting the vectors in a separate DB often does not makes. Just use Hana https://news.sap.com/2024/04/sap-hana-cloud-vector-engine-ai... ;-) IMHO putting the calculation of the embedding vectors into the db (even if it is just a remote call) is not a got idea. How do you react to failures of the remote call, security issues because of code running within your DB ..?

WASM ATM is IMHO most useful for VSCode Extension, where it can help to avoid the dependency nightmare that nodejs modules with native code cause.

I agree this is not a great test. What's good about it is that it is a constraint satisfaction problem, and I would expect LLMs to be pretty bad at unknown problems of this kind. Simple reason, an LLM only has a a finite number of layers and it cannot do arbitrary long searches.

Really nice"

What is the memory bandwidth? I could not find any numbers ..

It also very much depends on the quality of your documents. E.g. if you have a set of very good recipes, rather than a random set coming for the Internet, then you can get very good answers with RAG

Riding is more fun with less resistance, it just "feels better". That alone is an argument to choose the correct tire (there are trade offs obviously). From my own experience (MTB) I can say that the difference in rolling resistance are very noticeable.

Humans also stumble with that as well. Problems being CSV not really being that well defined and it is not clear to people how quoting needs to be done. The training set might not contain enough complex examples (newlines in values?)

https://eclipse.dev/mat/ can handle very large graphs of objects using a similar approach. It also does implement some kind of paging, such that you do not have to load the complete graph into memory when running some of the graph algorithms.

It is like others relying on tracking memory allocations. Which has it"s place, but to really analyze memory usage you would need to be able to analyze which "objects" are keeping other objects alive (dominator tree in MAT)

It does not really allow you to analyze what contributes to memory usage in a graph of "objects".

http://kohlerm.blogspot.com/2009/02/how-to-really-measure-me... do we really still not have a decent memory usage analysis tool for c/ C++ ?

Ok so Google is punished for being more open in the first case? Sounds fishy to me ...

What does thad mean "part of the OS"? It is not built into the kernel, and is therefore no part of the OS. I do not see why the same would not apply to Apple. But maybe the US legal system is just totally broken ...

If you want to run an LLM locally then an M2, otherwise this.