industrial technologist & digital innovator. data strategist with too many ideas, a liking for plants and a speaking computer which shares this affection. 🌸
@nathanrchn@liquidai I have big hopes for this. See I believe that these small models will solve very specific problems in processes which bigger models er too slow or expensive for. When it can solve small steps, it can solve everything (with good orchestration).
‼️ BREAKING: An active npm supply chain attack has compromised at least 868 packages carrying over 2 billion monthly installs with a credential-stealing worm. Shai-Hulud is back.
It started with the compromise of the GitHub account of the maintainer behind keyv, a library with roughly 127 million weekly npm downloads.
A preinstall hook fires on npm install and drops a stealer that sweeps npm, GitHub, AWS, Kubernetes and Vault secrets, and then spreads to more maintainers.
Finding a truly functional alternative to Windows Terminal is like naively opening Pandoras box. And I hate to admit it, that an actual Microsoft product is... eh... okay.
@solopribuilds Language models alone are a dead end and only a tool or stepstone when it comes to understanding reasoning. And for a 2nd one: model integrated test time reasoning is senseless bs.
@Skskp99@solopribuilds *won't eliminate all~ I'd say. Don't forget that we can differentiate between programmer & engineer. Some have personal emotionally biased opinions about AI. These people will be overrun.
Just reflect on the significance of this. Imagine how it was a year ago, two years, three and five years. Maybe Kurzweil was not wrong with his singularity? If we can achieve this, it means research is sped up too.
We ran a test between the new DeepSeek V4 Flash, Kimi K3 and GLM 5.2.
3 open models. 1 prompt. /design command
Reviewed gameplay features, UX/UI and cost.
🔹 Kimi K3 → 9.5/10 · $0.0740
🔹 GLM 5.2 → 9/10 · $0.0480
🔹 DeepSeek V4 Flash → 7/10 · $0.0005
DeepSeek V4 Flash is ~150x cheaper, but rough around edges.
Kimi K3 and GLM 5.2 built the most polished gaming experience.
@luckeyfaraday Don't say that so quickly. You have many factors: the way you work, the instruction format, the orchestration, the agent it runs in. All quality gates.