vladimirzaytsev
vladimirzaytsev on Hacker News
Re:Intercom changes name to Fin
on Wednesday May 13, 2026 @07:31PM (#48129057)
who cares?
Re:BYOMesh – New LoRa mesh radio offers 100x the bandwidth
on Sunday May 03, 2026 @09:03PM (#48003423)
It's clearly AI.
I’m struggling to see the value here. At $50, this seems hard to justify given the availability of cheaper, well-supported options like MuziWorks Duo, Ebyte, and other newer LR1121/LR2021-based designs. Those chips offer both 2.4 GHz and sub-GHz support in a single modern package, often at roughly half the price, which makes the SX1281 feel fairly outdated.
Re:BYOMesh – New LoRa mesh radio offers 100x the bandwidth
on Sunday May 03, 2026 @08:52PM (#48003340)
Nvm, looks like AI scam.
Re:BYOMesh – New LoRa mesh radio offers 100x the bandwidth
on Sunday May 03, 2026 @08:47PM (#48003305)
Sounds like bs. Why would someone pay $50 for almost 10 years old hardware when there are plenty of well-supported and cheaper options like MuziWorks Duo / Ebyte / etc with newer LR1121 or LR2021 which combine both 2.4G and SubG bands in single and modern chip at 1/2 of the cost less? SX1281 and SX1281 are relics.
Re:Show HN: Optimizing LiteLLM with Rust – When Expectations Meet Reality
on Tuesday November 18, 2025 @02:00PM (#45970456)
Is this whole post and github repo LLM-generated slop?
Re:Harder, Better, Faster, Stronger Version of Uber H3 in Rust
on Monday November 03, 2025 @04:28PM (#45804669)
H3 is preferred for geo analytics because it produces a more uniform spatial index with low distortion and consistent distances between cells
Its primary use case was efficient spatial aggregation for applications like pricing, demand forecasting, positioning etc.
Re:Embeddings are a good starting point for the AI curious app developer
on Wednesday April 17, 2024 @04:47PM (#40069900)
You may get way more accurate results from relatively small models as well as logits for each class if you ask one question per class instead.
Re:Embeddings are a good starting point for the AI curious app developer
on Wednesday April 17, 2024 @03:57PM (#40069416)
Likely not, embeddings are very crude. Embeddings of a text is just an average of "meanings" of words.
As is embeddings lack a lot of tricks that made transformers so efficient.
Re:ExecuTorch: Enabling On-Device interference for embedded devices
on Tuesday October 17, 2023 @03:30PM (#37920362)
This is very cool.