- The Prohuman
- Posts
- Google launches Gemini 4 Neon at a lower cost
Google launches Gemini 4 Neon at a lower cost
Plus: Anthropic's IPO reveals a major win for Amazon

Hello, Prohuman
Today, we will talk about these stories:
Gemini 4 Neon targets coding and complex AI tasks
Google will keep Gemini chats for two months
Amazon and Google will own 28.4% of Anthropic
The voice-to-text app that built its own speech model
Most dictation apps are a wrapper around someone else's speech model. Wispr Flow trained its own.
It's called Canto, and it was built for the way people actually talk to AI: thinking out loud, changing direction mid-sentence, half-whispering in an open office.
Wispr tested it on 10 hours of real dictations from more than 2,300 speakers. Canto had the lowest word error rate of every model tested, including models from Google, OpenAI, AssemblyAI and Deepgram. That means your prompts to ChatGPT, Claude and Cursor come through the way you said them.
Talk naturally in any app and Wispr Flow types clean, formatted text, 4x faster than typing.
Google's Gemini 4 Neon makes advanced AI cheaper to run

Image Credits: Google
Google says its new Gemini 4 Neon can match larger AI models on demanding tasks while running faster and costing less.
Released November 18, Neon targets software development, scientific research, and complex business work, with pricing starting at $3 per million input tokens and $15 per million output tokens.
Google says Neon delivered comparable quality to Gemini 4 Carbon at three times the speed and 60% of the cost in early evaluations by Glean.
That's worth watching.
I think the biggest opportunity is for companies running AI agents repeatedly throughout the day, where processing costs and execution speed have a direct impact on what they can afford to automate. Even small savings become significant when the same tasks run thousands of times.
Neon also supports a 1 million-token context window and can work across spreadsheets, PDFs, and code repositories.
The real test will come when developers start using it on everyday production tasks, where reliability matters as much as benchmark scores.
Will those reported efficiency gains hold up outside controlled evaluations?
Google is keeping Gemini chats longer, even with history off

Image Credits: Google Blog
Turning off Gemini's chat history won't stop Google from retaining your conversations for around two months.
Google is extending its retention period from 72 hours to roughly 60 days, citing safety concerns and the need to investigate potential misuse of its AI systems.
For comparison, OpenAI and Anthropic generally retain temporary chats for up to 30 days, although exceptions can apply.
That's a substantial increase.
Google says these retained conversations won't be used for AI training or product improvements. It's also separating chat history controls from AI training permissions, giving users more control over how their conversations are used.
I think separating those settings is a sensible move, especially for people who want to revisit previous chats without contributing their data to model training. But keeping conversations for two months raises reasonable questions about privacy.
Even when a chat disappears from your screen, the underlying data can remain on company servers.
Google says the extra time is necessary to investigate complex security threats, but it hasn't explained why the previous three-day window became insufficient.
How much retention is actually necessary for safety?
Amazon stands to gain twice from Anthropic's IPO

Amazon is getting a particularly good deal from Anthropic's IPO.
According to Anthropic's latest filing, Amazon will own 14.7% of the AI company, while Google will hold 13.7%, giving the two cloud providers a combined 28.4% stake.
There's more to Amazon's position. It still holds investment notes that could eventually increase its ownership to 23.8%, even after choosing to convert only part of them into shares.
The interesting part is how much Amazon stands to earn beyond its stock holdings.
Anthropic has also committed to spending more than $100 billion on AWS services over the next decade, giving Amazon a significant source of future revenue as Claude usage grows.
I think that makes Amazon's position especially valuable. It benefits from Anthropic's rising valuation while collecting revenue from the computing infrastructure Anthropic needs to operate.
The IPO documents make that relationship much clearer.
The bigger question is how much financial influence cloud providers will have over AI companies as their computing costs continue to rise.
Prohuman team
Covers emerging technology, AI models, and the people building the next layer of the internet. | ![]() Founder |
Writes about how new interfaces, reasoning models, and automation are reshaping human work. | ![]() Founder |
Free Guides
Explore our free guides and products to get into AI and master it.
All of them are free to access and would stay free for you.
Feeling generous?
You know someone who loves breakthroughs as much as you do.
Share The Prohuman itβs how smart people stay one update ahead.



