armcat

5 ms·

31.
by armcat 1mo ago
Same here. The absolute star in my team is the youngest. But! I think what's happening is that the distribution in this age group has shifted. Essentially the absolute best (high IQ, super ambitious, great communicator, cracked builder
32.
by armcat 1mo ago
It's interesting because a study here in Sweden [1] produces identical results. To quote from the study: "An event study documents an accelerating decline in employment of 22–25-year-olds in high-AI-exposure occupations, reaching
33.
by armcat 1mo ago
This is a great read and I can extend this argument really to any small business. Tax filing, VAT, bookkeeping, privacy, consumer, workplace and sectoral requirements frequently impose a disproportionately large cost on microbusinesses. EU
34.
by armcat 1mo ago
This is super interesting because I moved from mobile telephony into ML and data science, and information theory and working with data in statistically correct way was what helped me! This was 10 years ago though.
35.
by armcat 1mo ago
It's best not to take advice on direction of careers from anyone. It's better to find and work on things that interest you, and then take advice from people that are amazing in that specific field. In mid 2000s in Australia all th
36.
by armcat 1mo ago
This is a great read. One of the most requested AI features I’ve seen is PPTX slide generation that matches a corporate template and assets perfectly. Claude with skills comes close but not close enough. The amount of fiddling needed to adj
37.
by armcat 1mo ago
Having built my own voice assistant ( https://github.com/acatovic/ova ) and having tried many other services and models, I feel the real win is when this is on-device, and by "on-device" I mean being very inexp
38.
by armcat 1mo ago
So it's effectively on-par with GLM 5.2 and GPT 5.6 Luna?
39.
by armcat 1mo ago
I feel the single mention of `currentDateTime` against a large number of date mentions (I counted six) in the system prompt gets it confused sometimes when it comes to dates. I actually see the same effect when querying via API (when I appe
40.
HuggingFace is down due to global AWS outage(status.huggingface.co)
5 points by armcat 2mo ago|1 comments
41.
by armcat 3mo ago
This looks great, will try it out. I had a crazy idea of doing a full agentic demo video narration and cutting, so might try your tool with some tweaks.
42.
by armcat 3mo ago
TTS has come incredible long way, there are so many options. There is Kokoro of course, then there is Pocket TTS which is also a tiny 100M model that allows voice cloning. There is also Chatterbox Turbo, which is bit bigger but also allows
43.
by armcat 3mo ago
Does anyone know if this happens in the Claude desktop app?
44.
by armcat 3mo ago
I find it astounding that ppl still comment “it’s still behind” or “it’s not the best model”. Everything is about the harness. Even the big AI labs are focusing on managing agents - sandboxes, memory, context, skills, loops. With the right
45.
by armcat 3mo ago
I have the opposite experience with this, and I personally also default to simplest solutions at least as a baseline. However it’s important to distinguish between simple solutions that approximate the problem very well, to simple solutions
46.
by armcat 3mo ago
I mean it's always nice to play around with sLLM finetuning, but for practical purposes I would always start with a lazy learner using embeddings (something like a small Stella model), pre-embed the topics/categories, embed the qu
47.
by armcat 3mo ago
That movie has aged incredibly well!
48.
OpenAI to open office in Stockholm (Swedish)(efn.se)
4 points by armcat 3mo ago|0 comments
49.
by armcat 3mo ago
Neko Health has been doing this now for a few years. What I heard is that ultimately it doesn’t solve much (other than them privately collecting all your data) because there are lot of false positives and these false positives are deferred
50.
by armcat 3mo ago
Can this be used for ML/AI projects as well? I'm thinking for version controlling LoRa finetunes, finetuning data (which can consist of text, images and audio), safetensors, etc?
51.
by armcat 3mo ago
As an amateur watercolour artist (shameless plug: https://www.instagram.com/p/DBlKG5cMPxa ) I have to say the feeling your made with this wash is gorgeous. Back in the analogue world - paper grain and type/brand ha
52.
by armcat 3mo ago
I keep seeing these "sovereign" LMs time and time again. In Sweden we had GPT-SW3 ( https://www.ai.se/en/project/gpt-sw3 ) and same story there. Instead of burning money on "sovereign" claims, na
53.
Google AI Instruments (Magenta RealTime 2)(magenta.withgoogle.com)
3 points by armcat 4mo ago|0 comments
54.
Cosmos 3: Omnimodal World Models for Physical AI(research.nvidia.com)
4 points by armcat 4mo ago|0 comments
55.
by armcat 4mo ago
This looks beautiful and I'm sorry the current state of affairs has made you not want to publish the code, I would love to play around with it. Regarding your decision to build - I feel you, I've had the same happen to me for ever
56.
by armcat 4mo ago
How quickly they get new models supported on the API and it just works, is insane!
57.
Ask HN: What do we mean by compute scarcity?
1 points by armcat 4mo ago|2 comments
58.
Show HN: Simple Sprite Sheet Generation(github.com)
3 points by armcat 4mo ago|0 comments
59.
by armcat 4mo ago
This person did a great comparison against Qwen models, and despite them having 8x less active params, they outperform the Cohere model in every category: https://x.com/DJLougen/status/2057196012918149368?s=20
60.
by armcat 4mo ago
What an awesome story. Not too many stories about Aussies out there, but what Han brothers are doing with Unsloth in AI, and stories like this one, makes this fellow Aussie super proud!

More ›