Sponsored by
The way we talk to computers is shifting fast, from glasses that read silent speech to agents that run your errands. We checked the claims against the companies and the reporters who tested them, and flagged what is still a promise rather than a product.
100+ ChatGPT Prompts to Revolutionize Your Day
Supercharge your productivity with HubSpot's comprehensive guide to Chat GPT. This free resource is your fast track to AI mastery:
Industry-Specific Use Cases: 15+ real-world applications across various sectors
Productivity Guide: 21 best practices to 10x your efficiency with AI
Prompt Powerhouse: 100+ ready-to-use prompts for immediate implementation
Challenge Buster: Overcome common AI hurdles with expert strategies
Plus, in-depth sections on email composition, content creation, customer support, and data analysis.
TODAY IN AI
3 things that happened while you were busy
1. Glasses that turn silent mouthing into text opened for preorder.
A startup called Subvocal opened reservations for smart glasses that read the small muscle movements of silent speech using radar, with no camera and no sound, and type what you mouth. The early price is $499, the company raised $1.3 million from Y Combinator and Afore Capital, and a launch post drew about 7.3 million views. Read Digital Trends's report. The claimed 200-plus words a minute is unverified, the glasses ship in early 2027, and you still have to deliberately mouth each word.
2. OpenAI plans to buy Google's AI chips.
OpenAI reportedly plans to buy Google's TPU chips alongside its Nvidia hardware, and Oracle is said to be bringing TPUs to its cloud in 2027, which would loosen Nvidia's grip on AI computing. The reports surfaced in late October and neither company has confirmed the details. See Decrypt's report. More chip suppliers tends to mean cheaper compute, which eventually reaches the prices you pay to use AI.
3. Meta's personal AI agent reached more countries.
Meta expanded its Muse personal agent to Singapore, where it connects to Gmail, Google Calendar and Instagram and ties into local services like Shopee, Grab, Agoda and Klook, having already launched in the US, Canada and Japan. It asks for confirmation before sensitive actions like sending email or buying something, per a roundup of the week's releases. See the roundup. Shopping agents are moving from demos to everyday apps, one region at a time.
A CRM so smart, it updates itself.
HubSpot's CRM is so smart, it now updates itself. Calls get logged and summarized the moment they end. New leads get researched and a first outreach drafted, automatically. Deals move forward on their own, with next steps flagged before you have to ask.
Even the one repetitive task you've been meaning to fix can now run itself, no code required. Customers using it are already seeing the difference.
This isn't the CRM you have to work hard to figure out. It's the one already working before you log in.
TODAY'S TUTORIAL
Dictate instead of type, and let AI clean it up
Silent-speech glasses are still a year out, but you can get most of the speed today with the dictation already on your phone and computer, plus a quick AI cleanup. Here is how to draft three times faster without touching the keyboard. See where this is heading.
1. Turn on dictation. On iPhone and Android, tap the microphone on the keyboard. On Mac press the dictation key, on Windows press the Windows key plus H. It is free and built in.
2. Talk in full thoughts. Speak the whole message in one go without stopping to fix words. Say the punctuation out loud, like "period" or "new line," and ignore the small mistakes for now.
3. Paste it into an AI. Drop the raw text into Claude or ChatGPT and ask it to fix the grammar and tidy the flow without changing your meaning or your voice.
4. Save your cleanup prompt. Keep one reliable instruction (there is one in the Prompt Station below) so every dictated note gets the same polish in a single tap.
5. Use it where typing is slow. Reply to customers, draft product notes or capture ideas while walking. Dictation plus cleanup is fastest for long messages, not two-word replies.
Good to know: Phone dictation usually runs in the cloud, so avoid speaking passwords or card numbers into it. For private notes, check your device's on-device dictation setting.
FROM THE FRONTIER
The keyboard is slowly losing its job
The new input. For 40 years we have talked to computers by typing. Silent-speech glasses, better voice modes and agents that act on a sentence are all chipping at that, trying to make the keyboard one option among many. See the glasses.
The muscle behind it. All of this runs on expensive chips, which is why OpenAI reaching for Google's TPUs matters. More suppliers competing means the compute under these features gets cheaper over time. See the chip plan.
The hands doing the work. Once input is this easy, agents do the rest. Meta's Muse booking and buying through everyday apps shows where a spoken request is headed: not an answer, but an action. See the agent.
The reality check. The glasses are unproven and a year out, the chip deals are unconfirmed, and handing an agent your inbox and wallet raises real privacy questions. Early does not mean ready. See the caveats.
The reframe. Do not wait for the glasses. Get fast at voice and agents with the tools you already have, so when the hardware arrives you are improving a habit instead of starting one.
MONEY FORWARD
Three ways to turn this week's AI news into income
1. Voice-workflow setup for busy professionals. Plenty of people type slowly and never learned dictation. Offer a short paid setup: configure voice-to-text and a cleanup routine on their phone and laptop so they draft hands-free. See the trend driving it.
2. Review new AI gadgets for your audience. Viral devices like these glasses pull huge attention. Make honest demo and review content for a niche, and earn through affiliate links, sponsors or a paid community. See the attention these get.
3. Build shopping-agent-ready product pages. As agents like Muse start buying through apps, stores with clean titles, prices and stock data get picked. Offer to tidy a shop's product feed so agents can read and recommend it. See where agents are shopping.
GET THE EARNINGS GUIDE
Want the step-by-step version of ideas like these? Get The Aiwire earnings guide at theaiwire.co. These are ideas, not income promises, and results depend on your skills, market and effort.
IN THE KNOW
What people are actually watching and sharing
Google's strongest model. Google announced Gemini 4 Carbon, which it says tops the Artificial Analysis Intelligence Index, priced at $2 and $10 per million tokens, starting with select cyber defenders before a wider rollout. The benchmarks are Google's own. See the release.
Claude's plan shuffle. Anthropic renamed its Max 20x plan to Ultra at the same $200 a month and is deprecating Opus 4.8 fast mode on November 30, so heavy users should plan to move to Opus 5.5. See the change.
Gemini can run your screen. Google's Computer Use tool, which lets an AI operate a screen by clicking and typing, moved from preview to general availability for developers. See the details.
One quota for ChatGPT. OpenAI said that from December, Plus, Pro and Business plans will share a single token-based usage quota across ChatGPT and Codex, with quick chats not counting against it. See the plan.
Track your AI spend. Anthropic opened a beta Usage and Cost Analytics API for Claude Console and Enterprise accounts, letting teams break token spend down by model, workspace and user. See the tool.
TRENDING TOOLS
5 tools worth a look this week
1. GPT-6.1 Terra. OpenAI's cheaper mid-sized model for everyday computer work, at $1 per million input tokens and $5 output, available in the API, ChatGPT Work and Codex. Good for running automations on a budget. See the release.
2. Meta Muse. A free personal AI assistant on iOS, Android and the web that connects to your email, calendar and some shopping apps, and asks before it sends or buys. Now live in more countries. See the listing.
3. Gemini Computer Use. A developer tool from Google that lets an AI click and type on a real screen to finish tasks, now generally available through the Gemini API. Useful for building agents that use normal apps. Read the overview.
4. Mistral Large 4. An open-weight, roughly trillion-parameter model that is strong at coding and free to download, if you have the hardware to run it. Carried over from last week as a standout. Read the details.
5. ElevenLabs SFX v3. Generates sound effects from a text prompt and blends several into one clip, handy for short video and podcasts, with a free tier. Still one of the week's most useful creator tools. See the tool.
PROMPT STATION
Turn messy dictation into clean writing
This is the cleanup prompt from today's tutorial. It works in Claude or ChatGPT and fixes dictated or rambling text without flattening your voice. Save it and run every voice note through it.
COPY AND PASTE THIS PROMPT
Here is text I dictated by voice, so it has mistakes and no punctuation: [PASTE YOUR DICTATED TEXT]. Clean it up into [FORMAT, e.g. a short customer email / product description / set of notes]. Fix grammar, spelling and punctuation, and tidy the flow, but keep my meaning and my natural voice. Do not add facts I did not say or make it sound corporate. Return only the cleaned version, then list anything that was unclear so I can confirm it.
Tip: Swap [FORMAT] for what you are writing, like a WhatsApp reply, a blog intro or a to-do list. For a consistent voice, add one line of your own writing as an example and tell it to match that style.