Gemini 3.8 voice cloning: where it works and what consent covers
Google's new text-to-speech copies a voice from 30 seconds of audio. What one test cost, where Google blocks it, and what its consent check actually protects.
AI trend breakdowns
Every breakdown we have published, newest first.
Meta's Muse wants your work accounts. Decide what it can touch
Meta's Muse agent can now reach Notion, GitHub, Box and your email. What its safety promises cover, what critics found, and what to keep out of its reach.
Qwen-Image 2.1 says open-source; its licence says no paid work
Qwen calls its new image model open-source, but its licence allows research and evaluation only. What that means for client work, and your options.
Claude Opus 5.5: cheaper per token, unproven per task
Anthropic prices Opus 5.5 20% under Opus 5 and claims 40% less per task at its defaults. Nobody outside Anthropic has tested that. What breaks, and how to check.
DFlash 2: a 3.43× drafter that fades to 1.45× under load
Inco AI's block-diffusion drafter reaches 3.43× on GSM8K with one request in flight and 1.45× at concurrency 32. Here is what you pay in memory and variance.
Jev Ultrafast: the browser agent that stopped writing text
browser-use's jev-ultrafast finishes a Google Flights search in 7.1 seconds by replacing the agent loop with one typed decision per cycle. Here is the catch.
Laya on the Neural Engine: 4.98 ms and 0.154 J per decision
mizorewww's Core ML port runs one Laya typed decision in 4.98 ms and 0.154 J on an M3 Max. What the parity fixtures prove, and what they do not.
Nimble: 2,676 examples to get within three points of Jev
Bespoke Labs fine-tuned Qwen3.5-9B on 2,676 synthetic examples. It reaches 90.12% against Jev's 93.21%, and its evaluation covers six of ten trained domains.