Stories by soft-research

The Companies Cutting Headcount for AI Will Lose to the Ones Who Didn't

1024-bit prime,аvrg speed of~34ms;minimum latency:1.90 Ms.;over 1000 generations

Ultra-fast 1024-bit prime generator via Hilbert-Pólya spectral law

Ultrafast spectral primes generator (Python code)

5x faster prime generation than OpenSSL via Riemann zeta zeros (SpectralPrimes)

From bigger models to better intelligence:what NeurIPS25 tells us about progress

Grounding Text-to-Image Models for Controlled High-Quality Image Generation

Show HN: Text-to-Video Arena

Show HN: Open-Source Python REPL with AI Tutor for Learning and Problem-Solving

Show HN: How to guide on training Llama-405B using PyTorch distributed APIs

Show HN: I made an AI comparison tool

AI/LLM Aggregation Tool

Show HN: Rewriting My High School CS Teacher's 2D Graphics Library in JavaScript

Show HN: Mood Surf – A topic-based discovery engine for Twitter threads

What I learned from running a concierge search engine

1