HeyMeco

Much anticipated and finally available. I guess thank you Nvidia and regular installation on top of Debian Trixie works too so we should see support for SBCs and similar soon even though that would never be officially supported by them

This is the result of me being not happy with cli based serial consoles like tio, screen etc. When I debug many SBCs with different baud rates I just want a simple modern interface with a drop down menu for baudrate (accepts custom higher ones too). There is already a MacOS app that comes close to this but costs $39 and doesn’t have a modern terminal like ghostty behind it. This is free, will continue to improve and open source.

According to the dev who shared the screenshot of his RTX4060 ti running on a Radxa Orion O6 (Arm SystemReady) board they only needed to add the PCIe ID to the driver .inf file and it started to work.

The developer preview is intended for RTX Spark

Yet another local transcription tool but this uses Parakeet V3 (with FluidAudio / Apple CoreML under the hood) instead of Whispr and allows for easy diarization (speaker detection) and quality control of the result through an HTML interface.

This tool is a result of me realising there are no real options for interview transcriptions with speaker detection that can be used in an university study evaluation context where you do not upload the data into a someone’s else’s cloud with questionable privacy policies.

Feature summary:

- Diarized transcription — identifies speakers and assigns words to each speaker

- Plain-text output with timestamps and speaker labels

- Interactive HTML with: Color-coded speaker segments Word-level highlighting synced to playback - Click any word or segment to seek in the audio

- Rename speakers

- Built-in audio player (with playback speed selection)

Yeah fell into the same trap. Super annoying

They’re independent

Xcancel Link: https://xcancel.com/nvidia/status/2060390710797328574

Interesting takeaway:

"Was it faster than doing it solo? Probably yes, on the order of 3–5× for the work captured in these transcripts, with the speedup concentrated in:

Mass refactor / rename work (10× or more — Claude can rename a function and update 200 call sites in minutes). Code review breadth (subagent fan-out reads more files per hour than any human can). Performance analysis writeups (the “table of metrics + paragraph of prose” format is something Claude is genuinely good at and is tedious for humans)."

Jellyfin is so much ahead when it comes to hardware acceleration on different platforms that it is actually sad to see how Plex has not made any improvements in years

So today I found out the cheapest E27 smart bulb that is 1500lm bright and >90cri is actually an Ikea Kajplats Matter over Thread bulb that costs 9.99€

And it has one trick up its sleeve:

▶ If you turn it off and on 12x times it switches to a zigbee mode so I was able to integrate it into my existing HomeAssistant setup even easier

ML-1740 is supported by the driver available here: https://www.bchemnet.com/suldr/index.html Works on my arm64 sbc as a cups printer for another samsung printer

The non marketing fluff version of the press release can be found here: https://news.ycombinator.com/item?id=47506641

They actually said it’s coming in 2027 and ARM-AGI-3 is marked as future https://xcancel.com/mecoscorner/status/2036510613350785418

It really is a choice for Arm to use their 2023 based mobile X4 cores instead of their current C1 Ultras for this. Hopefully they step up quickly

I wonder what the people at Ampere are thinking right now

This is my condensed version for the SBCwiki documentation focused on the key facts without all the unnecessary marketing around it

Instant Grep 13ms vs 16.8s ripgrep seems very useful

What we knew so far about the Arduino Ventuno Q:

- CPU: Octa-core with 4x A78 and 4x A55 - GPU: Adreno A623 - NPU: 40 TOPS dense Int8 - Wifi 6 (2.4/5/6GHz) & 2.5GbE LAN - 3x MIPI cameras at once

What is new:

- GeekBench 6: On Par with QCS6490 used in Radxa Dragon Q6A which starts at $70

- 6x faster than Arduino Uno Q

- 1/2 the performance of IQ9 series (all big core)

- Supports AV1 decoding and H265 / H264 encoding

- Further elaborated in the article

In the past you had to run the benchmark via an x86 translation layer like Fex / Box64 to get the Vulkan GPU benchmark. This was not reliable and led to crashes on some platforms. Thankfully we can run it native now

I stumbled upon the VChance SU13 (here the SU13TO for the touch enabled variant) which advertises 100% P3, 4K resolution and a maximum brightness of 500 nits and this AliExpress find got me intrigued.

In the review I take a deep dive into calibrating this monitor and figuring out if it can keep its promise (spoiler it can but you have to know how).