Feed aggregator

Article URL: https://chromewebstore.google.com/detail/calendario/bnmmigoiecilbkipdngjofpbpaoofnoi

Comments URL: https://news.ycombinator.com/item?id=49860762

Points: 1

# Comments: 0

Article URL: https://github.com/thruwire/foreman

Comments URL: https://news.ycombinator.com/item?id=49860753

Points: 1

# Comments: 0

Article URL: https://chesscheatdetector.com/blog/how-to-tell-if-someone-is-cheating-on-chess-com/

Comments URL: https://news.ycombinator.com/item?id=49860750

Points: 1

# Comments: 0

Article URL: https://aatishb.com/blog/2026/ai-jagged-intelligence/

Comments URL: https://news.ycombinator.com/item?id=49860730

Points: 1

# Comments: 0

Article URL: https://playdaily.org/noillirk

Comments URL: https://news.ycombinator.com/item?id=49860722

Points: 4

# Comments: 1

Article URL: https://bsky.app/profile/julieslalonde.bsky.social/post/3mweffhm4tk2b

Comments URL: https://news.ycombinator.com/item?id=49860717

Points: 1

# Comments: 0

Article URL: https://www.brookings.edu/wp-content/uploads/2026/09/4c_Van-Nieuwerburgh.pdf

Comments URL: https://news.ycombinator.com/item?id=49860708

Points: 1

# Comments: 0

Article URL: https://www.ft.com/content/3f406fbe-b72e-488f-9975-5b94e95dfe32

Comments URL: https://news.ycombinator.com/item?id=49860704

Points: 1

# Comments: 0

Article URL: https://newsletter.semianalysis.com/p/clustermax-30-the-industry-standard

Comments URL: https://news.ycombinator.com/item?id=49860698

Points: 1

# Comments: 0

Article URL: https://bjorg.bjornroche.com/management/delegation-and-trust/

Comments URL: https://news.ycombinator.com/item?id=49860683

Points: 1

# Comments: 0

Article URL: https://apps.apple.com/gb/app/chains-messenger/id6763361983

Comments URL: https://news.ycombinator.com/item?id=49860669

Points: 1

# Comments: 0

Article URL: https://news.ufl.edu/2026/09/ai-powered-vehicles/

Comments URL: https://news.ycombinator.com/item?id=49860649

Points: 1

# Comments: 0

Article URL: https://xendo.bearblog.dev/the-wait-equation/

Comments URL: https://news.ycombinator.com/item?id=49860585

Points: 1

# Comments: 0

Article URL: https://twitter.com/dhh/status/2103880596841783698

Comments URL: https://news.ycombinator.com/item?id=49860579

Points: 5

# Comments: 1

The US and China agreed to set up a communication mechanism for artificial intelligence-related incidents.

The post China and US Agree to Establish AI Safety Channel and Continue Trade and Military Talks appeared first on SecurityWeek.

Article URL: https://theshadowedarchive.substack.com/p/an-existential-guide-to-forgiving

Comments URL: https://news.ycombinator.com/item?id=49858818

Points: 1

# Comments: 0

Article URL: https://github.com/jaredks/rumps

Comments URL: https://news.ycombinator.com/item?id=49858815

Points: 2

# Comments: 0

Article URL: https://www.aljazeera.com/economy/2026/9/25/japan-moves-to-tighten-rules-for-foreigners-throwing-futures-into-doubt

Comments URL: https://news.ycombinator.com/item?id=49858810

Points: 2

# Comments: 0

Article URL: https://shnatsel.github.io/state-of-simd-rust-2026/

Comments URL: https://news.ycombinator.com/item?id=49858799

Points: 1

# Comments: 0

I’ve been experimenting with a local approach to some of the classification tasks people are using Jev for. This approach uses text embeddings + logistic regression

On Banking77, which contains 77 categories of banking support questions, I get 94.25% using bge-large-en-v1.5 for embeddings, and 93.28% with all-MiniLM-L6-v2 (only the classifier gets trained, the embeddings model stays unchanged). For comparison:

Model Accuracy Size/training time - IntenDD (SOTA): 94.86% (~350M params, hours on GPU) - This script: 94.25% (642 KB classifier, 3s on CPU) - ModernBERT fine-tuned: 93.99% (149M params, minutes on GPU) - BERT fine-tuned: 92.76% (110M params, minutes on GPU) - Jev (zero-shot): 87.0% (closed API) - Laya (zero-shot): 42.5% (421M params) MiniLM’s encoder is roughly 91 MB in FP32, BGE is about 1.2 GB, the classifier and scaler are only 642 KB

The linked gist evaluates all 3,080 official test examples. Running python banking77_gist.py uses MiniLM by default, and you can choose the larger BGE encoder. Everything runs locally (although it needs to download the embedding model on the first run)

This is just a proof of concept, but it shows you can get very good results without needing GPUs, large amounts of data or hours of training, if you mostly just need classification

The direction I'm now exploring, is a system that handles familiar decisions with classifiers first, but asks an LLM for help when needed, then uses the LLM responses to automatically improve its local classifiers

What are you using Jev/Laya for?

Comments URL: https://news.ycombinator.com/item?id=49858795

Points: 1

# Comments: 0