diminish
Joined 16 years ago
Karma
4,710
Submissions
1,706
Member since
Nov 21
Recent submissions
Comment · 5 months ago
Can you pls link to them?
Comment · 8 months ago
one feels the llm wow moment whenever what they do on an area has been surpassed by an llm. newer versions of llms are probably trained by the feedback from developer code agent sessions; so this is probably why pro developers started to feel "wow" recently.
the real challenge will be in the frontier of the human knowledge and whether llms will be able to advance things forward or not.
ps1; i'm using 5.3/o4.6/k2.5/m2.5/glm5 and others daily for development - so my work has 1.5x intensified - i tackle increasingly harder problems but llms still really fail big in brand new challenges like i fail too. so i'm more alert than ever.
ps2: syntactical autocomplete used to write 80% of my code; now llms replaced autocomplete but at a semanticlevel; i think and LLM implements most of my actions like a cerebellum for muscle coordination; but sometimes teaching me new info from the net.
Comment · 1 year ago
Mini swe agent, as an academic tool, can be easily tested aimed to show the power of a simple idea against any LLM. You can go and test it with different LLMs. Tool calls didn't work fine with smaller LLM sizes usually. I don't see many viable alternatives less than 7GB, beyond Qwen3 4B for tool calling.
> right tools allow small models to perform better than undirected tool like bash to do everything.
Interesting enough the newer mini swe agent was refutation of this hypothesis for very large LLMs from the original swe agent paper (https://arxiv.org/pdf/2405.15793) assuming that specialized tools work better.
Comment · 1 year ago
Multi-core operations like compiling C/C++ could benefit.
Single thread performance of 16-core AMD Ryzen 9 9950X is only 1.8x of my poor and old laptop's 4-core i5 performance. https://www.cpubenchmark.net/compare/6211vs3830vs3947/AMD-Ry...
I'm waiting for >1024 core ARM desktops, with >1TB of unified gpu memory to be able to run some large LLMs with
Ping me when some builds this :)
Comment · 1 year ago
> sad to see lack of tools.
Lack of tools in mini-swe-agent is a feature. You can run it with any LLM no matter how big or small.
Comment · 1 year ago
Firefox's problem is Mozilla; too much ad money made them evil. copying every chrome feature to boost search ads by breaking usability :(
We need a new browser fork! Linux way of governance could be awesome; a star leader/programmer- with community and enterprise contributors.
/just switched to librewolf for the short term; after years on firefox since early 0.x days
Comment · 1 year ago
One can build a better container orchestration than kubernetes; things don't need to be that complex.
Comment · 2 years ago
We need >2000 km ranges in one charge of electric vehicles for widespread adoption.
Comment · 3 years ago
Does ctransformer (https://github.com/marella/ctransformers#supported-models) support running refact?
I see that model type "gpt_refact" in https://huggingface.co/smallcloudai/Refact-1_6B-fim/blob/mai...
Comment · 3 years ago
Controversially, agricultural societies late to industrialization, have more population now and are replacing early industrial societies due to lower birth rates in industrial workers.
So I doubt humans with AI will replace humans without.
Comment · 3 years ago
Well said. Openai is a promise unkept. Thanks StabilityAI for existing.
Comment · 3 years ago
Past discussions https://hn.algolia.com/?query=Crystal%20for%20Rubyists&type=...
Comment · 4 years ago
Twitter similarly is full of ads/phishing attacks disguising as some other brands, banks etc.
I've reported dozens, yet they never finish coming at me.