The Frontier - SiliconNoon (Page 14)

The Frontier

Today in AI — September 15, 2026

Tuesday's through-line was where the model lives. Shanghai AI Lab put a 744-billion-parameter agent on Hugging Face with no announcement, MediaTek and Honor put one inside the phone's silicon and operating system, and Google handed its own engineers a rival's coder. The infrastructure bill for all of it also came due: gas forecasts, a stacked data center, and a cloud region AWS says it cannot rebuild. Models & Research * Shanghai AI Laboratory posted a 744-billion-parameter agentic model wit

The Frontier

TabPFN-3.5 tops the tabular benchmarks with a thinking mode

Three stories about money and rules: the enterprise data model that just got a reasoning variant, the odd-couple summit asking Trump for a superintelligence treaty, and Washington's plan to make Mexico's AI hardware North American. Prior Labs shipped TabPFN-3.5, a tabular foundation model it says now leads both of the external benchmarks the field uses for real-world business data. The company reports the base model finishing about 150 Elo points ahead of the previous overall leader on BeyondAr

The Frontier

TypeSafe exits stealth with $40M for models that don't chat

Three stories about removing the expensive part: a model that skips strings, a search system that skips thinking at query time, and an agent economy whose only product so far is cold email. TypeSafe AI came out of stealth with $40 million in seed funding and Jev, a frontier model that refuses to generate a single string of prose. The round was led by DCVC; the founder is Diogo Almeida, a former OpenAI researcher who worked on the instruction-following methods behind ChatGPT and left convinced t

The Frontier

Gemini 3.8 Live tops voice benchmarks at a sixth of GPT-Live's price

Two launches Tuesday, one pattern: the fight has moved off raw model quality and onto the price of a unit of work — an hour of conversation, or a hairdresser's portfolio post. Google launched Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking on Tuesday, two voice-first models the company says top the realtime benchmarks while undercutting the closest rival voice models on cost. Both roll out starting September 15 through the Gemini API, Google AI Studio and Search Live, with enterprise acce

The Frontier

The closed-model premium buys 4.4 months — then it's lock-in

Mozilla's second State of Open Source AI report, out today, puts a price and a shelf life on the thing every enterprise AI budget this quarter is paying for. That is more useful than any leaderboard — and the report is honest about the number it can't measure. What Mozilla actually measured The capability gap between the strongest US closed frontier models and the best Chinese open-weights models is 4.4 months, according to the report published September 15. The evidence is specific: Moonshot

The Frontier

Honor ships the first phone with a system-level agent harness

Two stories from the Chinese AI beat this morning, one about where the agent runs and one about what it can simulate. Honor unveiled MagicOS 11 today with what it calls the industry's first commercially deployed system-level Agent Harness, and the numbers behind it are bigger than the label. Its YOYO assistant now strings together more than 100 steps in a single instruction, with roughly 700 system tools it can call — nearly double the previous set — and more than 500 third-party Skills already

The Frontier

StepFun ships five voice models and takes two global No. 1 spots

A Chinese lab just shipped the whole voice stack in one afternoon — and the pacing fight in Washington got a very public answer from Nvidia's CEO. StepFun released five StepAudio 3 models on the same day, and two of them now sit at the top of Artificial Analysis's speech leaderboards. StepAudio 3 Realtime, ASR, TTS, Gen and Music went live on StepFun's open platform Tuesday, covering everything from full-duplex conversation to music generation. The company claims 98.9 overall on the Artificial

The Frontier

AI 101 — What is backpropagation?

Backpropagation is the algorithm that lets a neural network learn from its mistakes: it measures how wrong the model's answer was, then works backwards through the network to figure out how much each individual connection contributed to that error. Once it knows who is responsible, the model nudges each connection a little in the direction that would have made the answer less wrong. Do that billions of times and you get a language model. Almost every claim you read about AI training — a model "

The Frontier

Frontier agents flunked two NeurIPS papers — graded by the authors

Two very different pictures of where agents actually are today: still failing at the top of AI research, and already handling real work inside a car cockpit and ByteDance's enterprise stack. A new study turned frontier agents loose on two unpublished NeurIPS 2026 submissions, gave each six days and thousands of dollars of compute, then had the papers' original authors grade the output. The agents completed all of the engineering without human help and still could not make substantial progress o

The Frontier

Meta's distilled byte models break through the token ceiling

Two research results and one big European round: distillation gets a smaller vocabulary, physical AI gets open hardware, and embedded-device security gets unicorn money. Meta FAIR found the way to break the distillation ceiling is to shrink the vocabulary to 256 bytes. A paper posted to arXiv on September 11 (authored with a University of Washington team) trained 1.28-billion-parameter byte-level students — models that predict raw bytes instead of tokenizer chunks — from larger teachers, and th

The Frontier

Shanghai AI Lab ships a 744B agentic model with no announcement

The lab that spent last week talking about science as the next programming just put a frontier-class agent on Hugging Face without a press release — and the weights, not the blog post, are the argument. Shanghai AI Laboratory has posted Atria Dawn Preview, a 744-billion-parameter Mixture-of-Experts agentic model, under the MIT license on Hugging Face and ModelScope — with no announcement, no launch post, and no paper. The model card is the only statement the lab has made: it describes a "previe

The Frontier

Sakana trains 1000-layer networks with no backpropagation

A paper out of Sakana AI this week takes the oldest objection to local learning and answers it where it bites hardest — depth. Meanwhile the slowdown debate got its most concrete moment yet, on speakerphone. Sakana AI's PC-ALM is the first layer-local training method shown to work on networks up to 1000 layers deep, according to work published by the lab and posted to arXiv. Predictive coding has long been the credible alternative to backpropagation: each layer nudges its own activity to shrink

See all