The Frontier

New models, papers, and benchmarks

AI's price crash is 13x a year — only 3x of it is better algorithms

The Frontier

AI's price crash is 13x a year — only 3x of it is better algorithms

Two studies put a number on the same thing: buying a fixed amount of AI capability is collapsing in price — and almost none of the collapse is the labs getting smarter. Epoch AI's "The Plunging Price of Thought" finds that reaching a fixed score on AI benchmarks has fallen about 47 percent per quarter since 2023 — roughly 13 times a year — and says no previous transformative technology declined that fast, running 4x faster than DNA sequencing, 6x than compute, 18x than lithium batteries and 54x

Real conversation broke 80% of the voice-agent speech models

The Frontier

Real conversation broke 80% of the voice-agent speech models

Voice agents are sold on benchmarks recorded in quiet rooms by one person at a time. A new one measured them on two people actually talking — and the leaderboard changed. Ocular AI published a speech-to-text benchmark this week showing that 12 of the 15 models it tested transcribed real two-person conversation worse than the clean public audio everyone else scores on, and four of them more than doubled their word error rate. The benchmark, called Converse-STT and run with evaluation startup Ce

NIO's MM-Future plans driving by imagining 64 futures at once

The Frontier

NIO's MM-Future plans driving by imagining 64 futures at once

A NIO paper turns consequence-aware planning into a benchmark score, Waymo stacks 271 million driverless miles behind its safety claim, and the AlphaFold Database quietly became pandemic infrastructure. Ren Shaoqing's name is on an autonomous-driving paper for the first time in a decade, and the method is the one the world-model field has been circling: let the car imagine many futures and pick the one it likes best. MM-Future, posted to arXiv on September 17 by a NIO-led team with Ren — a co-a

SeeAct AI: embodied AI's endgame is robots that retrain themselves

The Frontier

SeeAct AI: embodied AI's endgame is robots that retrain themselves

A Chinese robotics startup founded this year is arguing that the next leap in physical AI will not come from bigger models or more human demonstrations, but from letting robots generate and bank their own experience. Also today: iFlytek's industrial arm ships a small energy model built to run on the factory floor. Yao Mu, an assistant professor at Shanghai Jiao Tong University, has founded SeeAct AI to bet that embodied intelligence ends in reward-driven self-evolution rather than more training

arXiv just funded its independence — as AI floods its queue

The Frontier

arXiv just funded its independence — as AI floods its queue

arXiv just stopped being a university project. On September 23 the preprint server announced $17.2 million in multiyear commitments from the Simons Foundation International, XTX Markets, and the Siegel Family Endowment, funding its first years as an independent nonprofit. The timing is not a coincidence: the same platform now absorbs more than 28,000 submissions a month, and computer science — AI work above all — is nearly half of them. The money is the smallest part of this story; the load is

DeepMind says Gemini 4 is in post-training, eyeing an early launch

The Frontier

DeepMind says Gemini 4 is in post-training, eyeing an early launch

Google's next flagship is months out, not weeks — and the summit that has dominated the week conspicuously skips the chip question. Koray Kavukcuoglu, Google DeepMind's senior vice president, said Wednesday that Gemini 4 is "in the early phase of post-training" and that he hopes to release it "much earlier" than the end of the year. He was speaking at The Information's AI Agenda Live Summit, in his first interview since taking over the lab. The phrasing is the news: "post-training" is the stage

Google's Gemini 3.8 TTS clones a voice from 30 seconds of audio

The Frontier

Google's Gemini 3.8 TTS clones a voice from 30 seconds of audio

Google put two text-to-speech models into the Gemini API today and made voice cloning a standard developer feature. Also: Meta's Connect keynote landed three products, and Microsoft's new Gulf number is mostly an old one. Google released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS on Wednesday, and the headline capability is replication — building a synthetic voice from a reference recording as short as ten seconds, with Google's own marketing using thirty as the round number. The models

Apple's LensVLM-9B reads text as images at 4.3× compression

The Frontier

Apple's LensVLM-9B reads text as images at 4.3× compression

Apple quietly put a 9B vision-language model on Hugging Face this week that reads documents as pictures instead of tokens — and holds its accuracy down to a quarter of the original text. It is the most interesting long-context idea to come out of a big lab this month, and it landed with almost no announcement. Apple released LensVLM-9B on September 21 — weights under its own research licence, code under its sample-code licence, and a paper that has been sitting on arXiv since May. The pitch: st

Claude found a CRISPR-like enzyme system — and can't say what it does

The Frontier

Claude found a CRISPR-like enzyme system — and can't say what it does

Anthropic's lab published an unfinished result and named it anyway, while Google turned voice cloning into a line item you can price by the hour. Anthropic says one of its Claude agents autonomously spotted a previously uncharacterized enzyme system buried in a DNA database — and the company published the finding before it knows what the system does. The search took roughly 950 Claude agents 21 hours and 210 million tokens: they gathered more than 200,000 reverse transcriptases, picked out 3,50

Basecamp Research raises $140 million to train AI on the planet's DNA

The Frontier

Basecamp Research raises $140 million to train AI on the planet's DNA

Two frontier labs spent Tuesday undercutting each other's token prices. The round that might matter longer went to a company teaching AI to read DNA no public database has ever seen. Basecamp Research, the London startup building biological foundation models from wild DNA, has raised a $140 million Series C led by S32 — a round Reuters reports values the company at $800 million. Nvidia's NVentures, Anthropic's Anthology Fund, the NATO Innovation Fund, the Rockefeller Foundation, Redalpine and T

Alibaba's Qwen Audio 3.1: five voice models, prices cut up to 95%

The Frontier

Alibaba's Qwen Audio 3.1: five voice models, prices cut up to 95%

Alibaba's Qwen team shipped a complete audio stack today — and repriced it like it wants to own the category. Also in this brief: a Claude Code privacy quirk that got confirmed and fixed within hours, and AI agents getting a leash for the mainframe. Alibaba's Qwen team has turned its audio research into a full product line: Qwen-Audio 3.1 bundles five speech models covering recognition, synthesis and real-time conversation, with price cuts of up to 95%. The lineup spans a speech-recognition mod

MIT's AI barcode finds 'zombie cells' without destroying them

The Frontier

MIT's AI barcode finds 'zombie cells' without destroying them

Aging research has a measurement problem: the clearest markers of a senescent cell are only readable after the cell is gone. MIT and Harvard have now shown a way around it — and the answer looks like a barcode. A senescent cell stops dividing but refuses to die, and the immune system gets worse at clearing those cells out as we age. When they accumulate they are linked to osteoarthritis, type 2 diabetes, tissue degeneration and cancer — though senescence also does useful work in embryonic devel

See all