maxloh

27 ms·

91.
by maxloh 3mo ago
Also Goose from the Agentic AI Foundation (AAIF) (subsidy of the Linux Foundation). https://goose-docs.ai/
92.
by maxloh 3mo ago
I don't find a closed-source Chinese agent system trustworthy. It is essentially a black box with full user permissions, meaning you are just handing over your entire system to a Chinese-owned server. With OpenCode and its GLM provider
93.
by maxloh 3mo ago
I don't find a closed-source Chinese agent system trustworthy. It is essentially a black box with full user permissions, meaning you are just handing over your entire system to a Chinese-owned server. With OpenCode and its GLM provider
94.
by maxloh 3mo ago
According to CursorBench [0], Fable is the first runner-up, scoring 72.9% ($18.02, Max), while Opus 4.7 Max hits 64.8% ($11.02) and GPT-5.5 Extra High sits at 64.3% ($4.37). I bet most American companies would choose Fable over GPT-5.5. Emp
95.
by maxloh 4mo ago
I am still waiting for the Mojo language to be fully open source before diving into it. Another language I'm excited about is Carbon. I can't wait for an MVP to actually try it. We have so many new languages incubating in the last
96.
by maxloh 4mo ago
Open source is a good thing, but I don't think what you are proposing is accurate. A different way to frame this would be that those bugs would never be surfaced or exploited if the software were proprietary.
97.
by maxloh 4mo ago
That phrasing is just a way of lying to yourself. If Fable were released as open-weights, I doubt anyone would ever consider using GLM or any other models over it.
98.
by maxloh 4mo ago
Local AI will never be as powerful as cloud-based models, at least not in the foreseeable future. We’re talking about a difference between 7B and 750B+ parameters, a 100x scale gap. I think the trend will be running open-weight models with
99.
Tabularis: Open-source desktop SQL client your AI agent can use(tabularis.dev)
3 points by maxloh 4mo ago|0 comments
100.
by maxloh 4mo ago
Copyright protects the presentation of knowledge, not the knowledge itself, which is uncopyrightable in almost all jurisdictions. As long as the book was a legal copy, that is allowed legally.
101.
by maxloh 4mo ago
There have been too many ads for this site posted this month. 8 of your 10 recent comments are just site promotions, which is getting annoying. By the way, I don't really see the value here. Most of this could be vibe-coded and reimple
102.
by maxloh 4mo ago
> They achieve this by reselling capacity from pooled Claude Max accounts, payments fraud, and also reselling the model output & reasoning chains to various Chinese labs. Claude never provides the raw reasoning chain. What you see is
103.
by maxloh 4mo ago
I wonder how much of the layoffs were caused by their license change in 2021. They lost a lot of goodwill back then. Some of their potential customers migrated to OpenSearch and never looked back, even after they backed down and went open-s
104.
by maxloh 4mo ago
I don't get it. Qualcomm has almost no products in the high-end inference/training market. The industry standard is the NVIDIA Hopper H100/H200. What could they possibly get from acquiring Modular?
105.
by maxloh 4mo ago
This project is already highly vibe-coded. Claude is the second contributor of the repo.
106.
Keeping the Web Open and Private in the Bot Era(blog.mozilla.org)
3 points by maxloh 4mo ago|0 comments
107.
Cloudflare Collaborates with Leading Browsers to Develop Privacy-First Protocol(cloudflare.net)
9 points by maxloh 4mo ago|1 comments
108.
by maxloh 4mo ago
It sounds like GET with a body is just undefined behavior. Why not just standardize it? It seems to be a better way than adding a new method.
109.
Palmier Pro: macOS video editor built for AI(github.com)
4 points by maxloh 4mo ago|0 comments
110.
by maxloh 4mo ago
Nemotron is vastly different from standard open-weight models. Its entire training pipeline is open-sourced, while other vendors typically only release the model weights.
111.
by maxloh 4mo ago
Great to see more fully open LLMs. I think a problem with open-weight models is that while you can improve them, you are not going to create the next generation of LLMs by fine-tuning. We are at the mercy of frontier labs for access to SOTA
112.
by maxloh 4mo ago
Other fully open LLMs include Allen AI's OLMo 3.1 and MBZUAI's K2 Think V2, both of which have released their full training pipelines and datasets. Nvidia Nemotron is also an open training source model, though a portion of its dat
113.
by maxloh 4mo ago
That is a really terrible design. I dealt with a few instances of online ID verification recently, and in my experience, they don't close your application when your photo is not clear. They mark it as "awaiting customer response&q
114.
by maxloh 4mo ago
Related article: https://pierre.computer/writing/on-rendering-diffs
115.
DiffsHub: View code changes from any public GitHub diff with a freaking-fast UI
2 points by maxloh 4mo ago|1 comments
116.
by maxloh 4mo ago
Context: Raku was formerly Perl 6; it was renamed in October 2019 for compatibility reasons. > The major goal Wall suggested in his initial speech was the removal of historical warts. These included the confusion surrounding sigil usage
117.
Trump says Apple to partner with Intel on US chip design, production(reuters.com)
5 points by maxloh 4mo ago|0 comments
118.
by maxloh 4mo ago
Paper: https://arxiv.org/html/2503.01000v2
119.
Research finds MV3 does not affect ad-blocking or anti-tracking effectiveness(github.com)
4 points by maxloh 4mo ago|1 comments
120.
by maxloh 4mo ago
In 2018, Baker received $2,458,350 in compensation from Mozilla. In 2020, after returning to the position of CEO, Baker's salary was more than $3 million. In 2021, her salary rose again to more than $5.5 million, and again to over $6.9

More ›