Earlier quoted context omitted.
I rather think the GP was being sarcastic. At least, I hope they were.
[flagged]
Please don't engage in political/nationalistic battle on HN, and make an effort to observe all guidelines when participating here.
Tongyi DeepResearch – open-source 30B MoE Model that rivals OpenAI DeepResearch
151–156 of 156 posts
Re: Tongyi DeepResearch – open-source 30B MoE Model that rivals OpenAI DeepResearch
#152Re: Tongyi DeepResearch – open-source 30B MoE Model that rivals OpenAI DeepResearch
#153Earlier quoted context omitted.
> or if it takes a couple of months to fold these advantages back into the frontier models. Right now, I believe we're seeing that the big general-purpose models outperform approximately everything else. Special-purpose models (essentially: fine tunes) of smaller models make sense when you want to solve a specific task at lower cost/lower latency, and you transfer some/most of the abilities in that domain from a bigg…
> If/when frontier model development speed slows down You do not believe that this has already started? It seems to me that we’re well into a massive slowdown
It's hard for me to say. I don't think you know you're on the S-curve until after the fact.
On the one hand, most models are "good enough" for chatgpt-like usage, and there it's hard to see/feel generation-to-generation improvements. On the other hand, if you look at instruction following, dealing with long context windows, >200 tool call interactions while staying on track, there's still plenty of improvements to be had. So, hard to say where we are.
Re: Tongyi DeepResearch – open-source 30B MoE Model that rivals OpenAI DeepResearch
#154Earlier quoted context omitted.
I use a Macbook Pro with 128GB RAM "unified memory" that's available to both CPU and GPU. It's slower than a rented Nvidia GPU, but usable for all the models I've tried (even gpt-oss-120b), and works well in a coffee shop on battery and with no internet connection. I use Ollama to run the models, so can't run the latest until they are ported to the Ollama library. But I don't have much time for tinkering anyway, so I…
How's the battery holding up during vibe coding sessions or occasional LLM usage? I've been thinking about getting a MacBook or a laptop with a similar Ryzen chip specifically for that reason.
Currently I don't use vibe coding or even code assistants, so I can't speak to how the battery fares when doing that sort of thing. I don't know how much or how intensively they need to run the underlying LLMs.
For chatting with LLMs via ollama, I've seen total power usage go to about 50W (on an M3 Max) while the LLM is active, which is about 3x-4x power usage compared to just idling with browsers and editors open.
So I'd estimate about 2-3 hours of continuous LLM use on battery. Because I have enough RAM spare, at least there's no need to keep shutting down and reloading models.
I haven't really pushed it to find out how long they run on battery, as I haven't used LLMs all that much.
I'm more interested in the underlying operations of how they work, investigating novel model architectures and techniques, and optimising performance, than actually using them as an end user :-) Similar to how I enjoyed writing game engines more than playing games :-) Maybe I'll get into using them more in future.
Re: Tongyi DeepResearch – open-source 30B MoE Model that rivals OpenAI DeepResearch
#155Has anyone found these deep research tools useful? In my experience, they generate really bland reports don't go much further than summarization of what a search engine would return.
I almost exclusively use Deep Research as inputs to LLMs for deeper domain knowledge including frontier scientific theories etc.
Re: Tongyi DeepResearch – open-source 30B MoE Model that rivals OpenAI DeepResearch
#156Sunday morning, and I find myself wondering how the engineering tinkerer is supposed to best self-host these models? I'd love to load this up on the old 2080ti with 128gb of vram and play, even slowly. I'm curious what the current recommendation on that path looks like. Constraints are the fun part here. I know this isn't the 8x Blackwell Lamborghini, that's the point. :)
> I'd love to load this up on the old 2080ti with 128gb of vram and play, even slowly. I think you mean ram and no vram. AFAIK this is a 30b moe model with 3b active parameters. Comparable to the Qwen3 MOE model. If you do not expect 60 tps such models should run sufficiently fast. I run the Qwen3 MOE Model ( https://huggingface.co/unsloth/Qwen3-30B-A3B-GGUF/blob/main/... ) in 4-bit quantization on an 11 year old i5-…
> I get ~12 tps with 16k context
FWIW Ollama at its defaults with qwen3:30b-a3b has 256k context size and does ~27 tokens/sec on pure CPU on a $450 mini PC with AMD Ryzen 9 8945HS. Unless you need a room heater, that GPU isn't pulling its weight.