decide1000

7 ms·

1.
by decide1000 14d ago
In the Netherlands there is a computer museum with a huge collection. The computers can be used by anyone. Really a great experience for kids these days and adults. https://www.homecomputermuseum.nl/en/
2.
by decide1000 20d ago
I use Whisper in combination with Qwen3.8 27b to create Dutch subtitles for Russian videos. I follow Ukrainian channels about the war and I am now able to understand them. In my tests with other languages Voxtral was better then Whisper. Wh
3.
by decide1000 22d ago
I agree. A better approach would be to have everyone access the harness for free, make it open source, when more resources are needed move the workload to Perplexity for a fee.
4.
by decide1000 23d ago
Qwen3.8-27B-PrismaAQUA-5.5bit-vllm The output quality is higher. It's held at full precision (not quantized).
5.
by decide1000 23d ago
On the DGX I get 44.5 tokens per second (NVFP4). With 8 concurrent it's 241 t/s total. I am using the PrismaAQUA standard 9.7 t/s + Dflash2 30 t/s + torch-compile 37 t/s c8 = 177 t/s
6.
by decide1000 23d ago
This should have been the link! https://pplx-dgx-spark-lp-demo.pplx.app/ Looks very useful update: It advertises with local harness and models but its closed source and you need a pro or max account
7.
by decide1000 27d ago
It creates a hyper personal daily podcast with latest niche news which is interesting for me alone. And it does that pretty well. You can probably create it yourself using an LLM, search and Elevenlabs.
8.
by decide1000 27d ago
Intel did some promising deals (dc, Europe, arc gpu) but still no real answers to AMD and Nvidia. I pulled out from Nvidia, MSFT and Intel and ramp up AMD. Too much happening for me in the AI space and miss the base that Intel used to have.
9.
by decide1000 27d ago
I use it for a while now. One thing that is super frustrating is that it saves it data (prompts etc) to the cache folder instead of applications data. Cache obviously can be deleted by the OS when disk space is limited or when there is a cl
10.
by decide1000 27d ago
I wouldn't describe it as luck, it's politics. The mother of Bill Gates had access to the board of IBM. She was also the co-founder of Microsoft.
11.
by decide1000 1mo ago
That is just an amazing website and initiative. https://rvembedded.com/products/project-lab/ Definitely interested!
12.
by decide1000 1mo ago
It's not just the AI. The post mentions poor choices because of a product management mindset instead of engineers leading the PMs.
13.
by decide1000 1mo ago
Great read. Exact pinpointing of my feelings about CF lately.
14.
by decide1000 3mo ago
A lot of replies here are about Mac devices and their support for these 27B models. I own a MacBook but use a Lenovo Thinkstation PGX to run my models. It has a gb10 Blackwell gpu and 128gb unified memory. You can connect multiple ones.
15.
by decide1000 3mo ago
If this is real I will stop my monthly subscriptions.
16.
Fable 5 wrote a Windows kernel in 38 minutes(tolmo.com)
6 points by decide1000 3mo ago|1 comments
17.
by decide1000 3mo ago
Finally! Now it becomes economic for us to make the move! Goodbye CloudFlare!
18.
by decide1000 4mo ago
As a Nvidia Spark owner I can confirm that Linux is much more performant then Windows on that chip. If the goal is "doing stuff with AI" Windows is the last OS you want to look at.
19.
by decide1000 5mo ago
Isn't voxtral much better?
20.
by decide1000 5mo ago
Mostly ecommerce and pricing data. I work for marketplaces, brands, retail stores and even our own saas competitors. We match the EAN (gtin) to the correct SKU within seconds (Google Shopping, Amazon, etc). Part of it is our own trained ML
21.
by decide1000 5mo ago
Despite that I find the goal of what you are trying to achieve questionable, I believe it should not be the AI that judges you here. We are all witnessing the start of an AI era that will not end soon. Guiderails are a part in this developm
22.
by decide1000 5mo ago
With scraper tech I mean a rust binary that is able to download and process thousands concurrent urls (millions per hour). Not to the same domain obviously. Paying more is not the issue here, its more the idea that an AI decides on what par
23.
by decide1000 5mo ago
I am not using brew. Just checked and it still says 2.1.98. Will try manual update.
24.
by decide1000 5mo ago
I am on the latest version available for me 2.1.98
25.
Claude Code Opus 4.7 keeps checking on malware
69 points by decide1000 5mo ago|63 comments
26.
by decide1000 5mo ago
I built a distributed DuckDB setup using OpenRaft for state replication. Every node holds a full copy of the database. Writes go through Raft consensus, reads are local. It's more like etcd-with-DuckDB than MotherDuck-lite. OpenDuck ta
27.
by decide1000 5mo ago
I'm building a Redis real-time backup platform with PITR (point in time recovery) to the exact millisecond. Besides downloads it has an emergency recovery option where you connect the Redis client directly to the master-replica to rece
28.
by decide1000 8mo ago
I think this model proves it's very efficient and accurate.
29.
by decide1000 8mo ago
I am on Linux for 26 years. Last 5 years I run PopOs! on my desktop. User friendly and stable, Ubuntu based.
30.
by decide1000 9mo ago
China and Europe (Mistral) show that models can be very good and much smaller then the current Chatgpt's/Claudes from this world. The US models are still the best, but for how long? And at what cost? It's great to work daily

More ›