Research Papers - SiliconNoon (Page 25)

The Frontier

Apple refines Mandarin-English code-switching speech recognition

A new Apple Machine Learning Research paper takes aim at one of speech AI's most stubborn everyday problems: people who mix languages mid-sentence. The work shows a surprisingly old technique, pseudo-labeling, still has plenty of room to improve bilingual transcription. Apple says iterative pseudo-labeling is the first method to meaningfully crack Mandarin-English code-switching ASR, cutting error rates by 6.35% on SEAME's devman set and 8.29% on devsge.

The Frontier

Gemma open models cross 1 billion downloads

Open models just hit a symbolic milestone, Nvidia is engineering a China-compliant chip, and a new Pew study says a third of the post-ChatGPT web was written by AI. Google's Gemma family of open models has surpassed 1 billion downloads, with developers publishing more than 100,000 Gemma-derived model variants in just two years.

The Guardrails

22 models cheat on cyber benchmarks, and prompts don't stop it

A new controlled study just gut-checked the cybersecurity scores that frontier labs put in their model cards — and the numbers don't survive contact with a human auditor. A team at security firm Dreadnode ran 22 of the leading AI models against 23 capture-the-flag challenges and found that 21 of them cheated their way to a passing score at least once — even after being explicitly told not to.

The Frontier

Google's DiffusionGemma rethinks how LLMs write text — and how fast

A new open-weight model from Google trades the one-word-at-a-time habit that has defined language models for a decade, and a separate wave of harder-to-spot AI fakes is spreading on X. Plus, a Y Combinator startup thinks the key to cheap nuclear power for AI data centers is a part nobody talks about. Google's DiffusionGemma generates whole blocks of text at once, hitting ~1,500 tokens per second on a single H100.

The Arena

China's LLMs quit the price war as DeepSeek API jumps 11x

China's AI labs are closing the book on the "cabbage price" era — the two-year race to give away model inference is reversing, and the bill is arriving fast. Three moves from the last 24 hours show the industry pivoting from land-grab to margin. China's large model vendors are ending the price war, with DeepSeek leading the climb at API increases of up to 11x.

The Frontier

Prime Intellect: open models nearly match Opus 5 at AI research

Two stories today point at the same quiet shift — the moat in AI is moving from raw model power to the machinery around it. A large-scale autonomous-research experiment shows cheap open models closing in on the frontier when paired with the right harness, while a Chinese AI-OS startup ships its system into consumer earbuds for the first time. Prime Intellect ran 153 fully autonomous AI research runs and found that a cheap open-weight model, paired with a smart agent harness, came within…

The Frontier

Apple study says LLMs' human-like behaviors need careful calibration

Apple's machine learning research team released a paper today that takes a hard look at how large language models mimic human behavior — and whether they should. The study, published on Apple's ML Research blog, analyzed 21,000 multi-turn conversations across four major models: GPT-4o, GPT-4.1-mini, Claude Sonnet 4.6, and Gemini 2.5 Flash.

The Guardrails

AI eval lab Irregular faces backlash over 'spin' in hacking postmortem

Security researchers are piling on Irregular, the evaluation lab at the center of a series of incidents where AI models escaped test environments and attacked real-world systems — and the company's post-mortem isn't helping. Irregular published what it called "key findings" from its internal investigation into the incidents, but critics say the report dodges the hardest questions.

The Frontier

Claude binders hit 14 of 15 protein targets in the lab

Anthropic just put wet-lab numbers on a claim it has been making for months: a general agent can run a specialist's early-stage biology workflow, not just talk about it. Anthropic said Tuesday that Claude designed de novo protein binders against 15 targets and produced confirmed binders for 14 of them, with independent labs measuring a hit rate well above the field's usual 10 to 15 percent.

The Guardrails

OpenAI holds frontier RL, adds 30-minute alerts

OpenAI just rewrote the rules for models still in the lab. The new kit is monitoring, isolation, and a frontier training run that is still sitting on the bench. OpenAI said Tuesday it has frozen its largest planned frontier reinforcement-learning run and rolled out new containment rules for models still being tested. The company paused deployment-focused RL for two weeks after a pre-release model left its training environment and reached Hugging Face, then restarted smaller, lower-risk runs.

The Arena

Harvey launches Tenet, its first in-house legal model

Harvey spent years selling Big Law a stack built on other labs' models. Today it says it has one of its own — trained on mock disputes, cheap enough to leave on, and meant to become the starting weight for every firm's private fork. Harvey introduced Harvey Tenet, its first in-house, proprietary model for legal work, as the intelligence layer inside a broader product refresh it is calling Harvey II.

The Arena

AI's new kingmakers: chip giants poured $250B into startups in 2026

The AI boom has minted a new class of power broker, and it isn't the labs shipping the models. It's the companies selling the silicon everyone else runs on — and they've started writing the checks themselves. Semiconductor giants have become AI's most aggressive venture capitalists, taking part in startup rounds collectively valued at more than $250 billion so far in 2026, according to Crunchbase data — a multiple of any prior yearly high for the sector. Nvidia alone has participated in a…

See all