analog chips could also be very interessting instead of using digital signals and processing them against the weights in the ROM. I have no idea if that scales with such big models though.
some years ago i tried this setup for a german company with a special char in its name („ä“) and failed because Apple was not able to match it against DUNS. It took months of support to get it done.
ef core is great for simple queries and modification of your data while using the changetracker.
You can use AsNoTracking/Projection to perform similar like dapper for queries.
When using command query seperation you can also use dapper for queries and ef core for commands.
in many cases: no service health alerts, no status page updates and no confirmations from the support team in tickets.
still we can confirm these issues from different customers accross europe. Mostly the issues are regional dependent.
what is a unhealthy request? is searching for a user which was _not found_ by the server unhealthy? was the request successful? thats where different opinions exist.
The idea is to replicate the weights of the network in the electronics. Somehow like our brains work? This way an analog input signal could lead to a neural network processed output signal without the digital emulation on an gpu. As this is very much simplified, the question is if this could work for modern llms?
I did!
with their own new hardware (https://www.home-assistant.io/voice-pe/)
sadly the microphones are way worse than e.g. in an alexa speaker.
Also the performance of the „voice pipelines“ (stt, llm, tts) are a bit of a pain because they are all in sequence and not e.g. using stream features.
isn’t a LLM itself basically a compression of the texts from the internet? you can download the model and decompress the (larger) content with compute power (lossy)
at a discussion about this issue i read a comment which i cant find right now but was very beautiful. It was saying that we should keep easter eggs like this, because in the end this is what differs us humans from machines.
In soccer they actual are. the ads are injected to the sideboards beside the playing field, so that if you watch the same game on different channels, they have different ads depending on their avg viewers.
how do you handle resetting a sql database after every integration test? Testcontainers may help here by spinning up a new instance for every test but that seems very slow
What I struggle with solutions like this is permissions. When having a permission system like row level security in place for sql, how to rewrite all this logic for the search queries?