roger_

I’m interested in using MCP for pair programming.

What tool can I use to point it to a directory and give Claude access to the code? I don’t want to have to write my own server.

Why monstrosity? I've not had any issues and their docker container is quite convenient.

Yeah but getting stuck on an old version of ESP-IDF is annoying.

I guess it could either mean “the patch will remove access to those commands” or “the patch will add the ability to disable those commands”.

Yeah but:

> Espressif will provide a fix that removes access to these HCI debug commands

Very annoying for researchers and tinkerers who would like to play with those commands.

I hate sensational stories like this. Now Espressif is gonna feel pressured to be even more closed.

Can anyone summarize the novelty here (for general time series)? Have these features been referenced recently?

I really hope they:

1) have a solution for spam (both blatant and subtle)

2) actively support their moderators

3) have well thought out and consistently enforced policies.

Reddit lost all of my goodwill years ago because of this.

Did Slashdot really kill itself or was it just never ambitious enough to seek a broader base or try to foster communities?

I don’t recall any user hostile practices, though I was just a passive user back in the day. I remember a few redesigns that people weren’t entirely pleased with, but nothing on the scale of the “new” Digg.

Basic math problems should be treated as a special vs general tool use. Why use Python when the computer evaluating the LLM can do math itself?

An aside: please use proper capitalization. With this article I found myself backtracking thinking I’d missed a word, which was very annoying. Not sure what the authors intention was with that decision but please reconsider.

I moved to AGH a while ago too.

Is there anything in Pi-Hole v6 that would make someone switch back?

This looks incredible, but I’m not a fan of the subscription pricing.

How about a hobbyist rate at least?

I don't think you can trigger by a label/area and also retrieve which device was the trigger (groups had that issue).

Organizing devices and creating automations is still very tedious.

I'd love to be able to add a device and describe it (where it is, what it does, etc.) and have HA automatically integrate it with existing automations or fuse it with other sensors. Maybe leveraging LLMs for this.

e.g.:

I buy a new leak sensor and add it to HA. I should just tell HA it's a leak sensor and it's in my laundry room and have it create an automation to send alerts, etc. when there's a leak.

Or I add a temperature sensor in my living room and have it automatically be fused with other sensors to update my living room average temperate.

This articles seems like it was at least partially written by AI. Lots of fluff and no clear explanation of what PINNs are and how they work (other than the code).

I’ve done simple one off web stuff over the years and yeah it seems much more straightforward nowadays.

Copying and undercutting someone else is certainly easier!

I’ve been wanting to do this for years but I’m still trying to figure out how to get started.

My skills are more in algorithm development (statistical signal processing, machine learning) and electronics than web coding though, so it’s probably not as easy as just making a simple website that does something slightly useful.

Interesting, I’ve always wondered if the assignment problem could be solved as continuous optimization. I’m familiar with manifold optimization but didn’t realize the permutation matrix could be expressed as a manifold.

Has this been applied to networks like DETR for object detection? I’ve never been clear on how they made the Hungarian algorithm differentiable (and it seems like their approach has issues).

Wonder if there'll be a day when a hobbyist can buy a < $1k SEM?

Other equipment like spectrum analyzers, VNAs, scopes, etc. are now way more accessible, but they seem way less complex.

Hope someone integrates this with Home Assistant soon!

This solves one of my biggest issues with AirTags and I'd be tempted to buy one if AirTag 2s weren't right around the corner...

> I wouldn't be confident that the Airtag protocol will be functioning in 2035

I'd agree if it were any company other than Apple. And if Apple goes under by 2035 then AirTags will be the least of our concern.

"A more intuitive way to express the updated state estimate..."

Interesting, thanks for sharing.

One question: what load is the matching network designed for? Did the designer find the equivalent small signal impedance of the diode network via simulation? Is the SPICE model even valid at 2.4 GHz? Is small signal even applicable?

It’s actually shown in the Wikipedia article: https://en.wikipedia.org/wiki/Kalman_filter#Details

A normal EWMA filter is a special case of a Kalman filter and I bet this could be re-cast as one too. Might need an extra state or adaptive process noise though.

I'd love to see SSMs replace transformers but adapting them to non-causal, 2D+ inputs doesn't seem that straightforward.

Is there a non-autoregressive future?

Great work, looks useful!

Any chance you can add a rational function version?

Came to suggest the same thing.

I’d love to be able to filter by CPU and generation.

Maybe using an LLM could help with the parsing?