In partnership with
|
BUILDER HUSTLE
AI · TECH · STARTUPS
|
|
|
In This Issue
|
|
🎨 Meta Launches Muse Image and Previews Muse Video
|
|
💻 Claude Cowork Now Follows You From Desktop to Phone
|
|
🗣️ Pocket TTS Clones Voices From Just 5 Seconds of Audio
|
|
🎁 Anthropic Gives Open Source Maintainers 6 Months of Claude Max Free
|
|
|
🎨 META · MUSE IMAGE & VIDEO · LAUNCHED
|
|
Meta's Muse Image Acts More Like an Agent Than a Prompt Box.
|
|
|
Meta Superintelligence Labs launched Muse Image, its first image generation model, and it works less like a typical prompt to image tool and more like an agent that invokes search and coding tools, self refines its own outputs, and can use extra compute at inference time for better results. It supports composing images from multiple references, editing with precision across several turns, and pulling in social context from Instagram, though that last feature has already drawn privacy criticism since it is on by default. Meta also previewed Muse Video, built on the same foundation, promising strong visual fidelity with native audio support, though it remains in preview for now. Muse Image is live today in the Meta AI app, on Instagram Stories, and on WhatsApp in select countries.
|
|
|
📌 Framing image generation as an agentic process, with self refinement and tool use, is a meaningfully different approach than a single-pass diffusion model.
|
|
⚡ Pulling social context from public Instagram accounts by default is the part likely to generate the most pushback and follow-up scrutiny.
|
|
|
Read Meta's Muse Image announcement →
|
|
How owning AI deployment expands your career
Across product, ops, and CX teams, a new kind of role is taking shape: the person responsible for making AI actually work, day to day.
On July 16, three people living this shift join a live roundtable: Simone Santiago Broad (Yoco), Yelva Espinoza (Zumba Fitness), and Fin's Dave Lynch. You'll hear what the job really looks like across industries, how they carved out these roles, the skills they'd hire for, and the challenges they're tackling now. Bring your questions, since the best moments happen live.
Register for the roundtable to save your spot.
|
💻 ANTHROPIC · CLAUDE COWORK · CROSS-DEVICE
|
|
Claude Cowork Keeps Working Even After You Close Your Laptop.
|
|
|
Anthropic is rolling out Claude Cowork on web and mobile, so a task you start on your desktop can be picked up and finished on your phone. Scheduled tasks now run without any device needing to stay online, so something like a Monday morning client prep can run at 6am and be waiting for you by the time you check your phone. Cowork works across your files, calendar, email, messaging apps, and connected tools until a job is done, and chat and Cowork are being unified into a single interface rather than two separate modes.
|
|
📌 Tasks continuing to run with your computer off is the real unlock here, turning Claude into more of a background worker than a chat window.
|
|
|
Read about Claude Cowork on web and mobile →
|
|
|
🗣️ KYUTAI LABS · POCKET TTS · OPEN SOURCE
|
|
A 100M Parameter TTS Model That Clones a Voice in 5 Seconds, No GPU Needed.
|
|
|
Kyutai Labs released Pocket TTS, a tiny 100 million parameter text-to-speech model that runs entirely on CPU, no GPU required, using a continuous audio modeling approach instead of fixed symbolic tokens. It runs about six times faster than real time on a Mac, with roughly 200 milliseconds of latency to the first audio chunk, and it can clone a voice convincingly from just 5 seconds of reference audio. It supports six languages out of the box, is fully open source under the MIT license, and was trained exclusively on public data, which makes it a strong option for local, private, or edge deployments where sending audio to the cloud is not an option.
|
|
📌 Running comfortably on CPU at six times real time speed removes the GPU requirement that keeps most decent TTS models out of reach for edge devices.
|
|
⚡ Cloning a voice from just 5 seconds of audio is a very low bar to clear, which makes personalization far more accessible than older TTS pipelines.
|
|
🔭 Being fully open source and trained only on public data makes this a clean option for privacy-focused apps that cannot rely on cloud TTS services.
|
|
|
Read the Pocket TTS technical report →
|
|
The Ultimate Guide for Usage-Based Pricing for SaaS and AI
Implementing usage-based pricing successfully requires more than just a pricing strategy.
Download this guide for practical advice and best practices when considering usage-based pricing.
|
🎁 ANTHROPIC · CLAUDE FOR OSS · FREE ACCESS
|
|
Anthropic Is Giving Open Source Maintainers 6 Months of Claude Max, Free.
|
|
|
Anthropic expanded its Claude for Open Source program to offer 6 months of Claude Max 20x completely free to eligible maintainers and contributors. That includes people who maintain critical packages, land meaningful pull requests, or serve as core contributors on widely used open source projects.
|
|
📌 Claude Max 20x is a high tier plan, so giving it away free for 6 months is a meaningful gesture rather than a token discount.
|
|
|
Apply for Claude for Open Source →
|
|
Stay sharp,
Builder Hustle
