In partnership with
|
BUILDER HUSTLE
AI · TECH · STARTUPS
|
|
|
In This Issue
|
|
🕶️ Linux AI Agents - On Your Face
|
|
🔀 Perplexity Splits AI: Local + Cloud
|
|
☁️ Google's Personalized Daily AI App
|
|
🎨 Reve 2.0 Turns Images Into Code
|
|
💻 Gemma 4 12B Runs on Your Laptop
|
|
🎙️ Free TTS That Clones Your Voice in 10s
|
|
|
🕶️ MONAKO GLASS · WEARABLE LINUX · AI AGENTS
|
|
The World's First Linux Computer You Wear on Your Face.
|
|
Monako Glass runs Buildroot Linux on an Arm Cortex A7 with a 0.5 TOPS NPU, letting you run Claude Code, Codex, Cursor, or any terminal-based agent directly in your field of vision. Battery gives you 4 hours screen-on or 8 hours normal use. Prescription lens slot included. The demo videos show vibe-coding while walking.. terminal HUD floating over reality. Pre-orders are open with August shipping targeted. Cyberpunk stopped being fiction.
|
|
📌 Runs any terminal-based coding agent - Claude Code, Codex, Cursor, Hermes - directly on the glasses.
|
|
⚡ Hand tracking + YOLO object detection baked into the NPU. No phone or cloud tether required.
|
|
🔭 Prototypes evolved from hand-made to CNC; injection molding is next. This thing is shipping, not vaporware.
|
|
|
Pre-order Monako Glass →
|
|
|
☁️ DREAMBEANS · GOOGLE LABS · PERSONAL INTELLIGENCE
|
|
Google Turns Your Activity Into a Daily Personal Feed.
|
|
Google Labs launched Dreambeans, a mobile app powered by Personal Intelligence that connects to your Google apps and builds a personalized daily story feed. It surfaces things you may have missed and deep dive topics tailored to your interests. It is currently invite only for US based Google AI Ultra subscribers aged 18 and above, with an open waitlist.
|
|
📌 Connects across Google apps such as Gmail, Calendar, and Search history to build a personal context layer.
|
|
⚡ Currently limited to Google AI Ultra users in the United States who are at least 18 years old. The waitlist is open.
|
|
🔭 The real test is whether Google can make a personal AI that feels helpful and not creepy. Early users will answer that quickly.
|
|
|
Join the waitlist →
|
|
Are you running your business on incomplete numbers?
Most small business owners have financials, but few have financial clarity. There's a real difference between books that are technically up to date and books that actually tell you what's going on in your business right now. When accounting is reactive — updated when there's time, reviewed at tax season — you lose visibility exactly when you need it most. You can't tell which clients are truly profitable. You can't spot a cash flow gap before it becomes a crisis. BELAY's outsourced accounting team changes that.
|
🎨 REVE 2.0 · 4K IMAGE MODEL · LAYOUT AS CODE
|
|
Reve 2.0 Treats Every Image as Editable Code and Ranks Second Worldwide.
|
|
Reve 2.0 is a 4K image model that generates each image as segmented and labeled layout code. Every region is addressable, editable, and precisely controllable. It reached the number two spot on the Arena Text to Image leaderboard, ahead of Nano Banana 2 and GPT Image 1.5. The layout bet gives you pixel level regional control instead of prompt roulette.
|
|
📌 Every generated image is structured layout code, so you can edit individual regions without regenerating the full image.
|
|
⚡ Ranked second on the Arena Text to Image leaderboard globally and ahead of GPT Image 1.5. Built by a relatively small team.
|
|
🔭 Images as code is the direction design tools are moving toward. Reve is one of the first models to actually ship this idea.
|
|
|
Try Reve 2.0 →
Read The Layout Bet →
|
|
|
💻 GEMMA 4 12B · GOOGLE · LOCAL MULTIMODAL
|
|
Gemma 4 12B Brings Text, Vision, and Audio to a 16 GB Laptop.
|
|
Gemma 4 12B is a unified encoder free multimodal model that handles text, vision, and audio in one weight file under the Apache 2.0 license. It runs on just 16 GB of VRAM or unified memory and reaches performance close to the larger 26B model while using less than half of the memory. Vision uses a lightweight 35 million parameter embedder instead of a separate frozen vision transformer. Audio projects raw signals directly into the LLM token space. The context window is 256K tokens. It already ships with support in Hugging Face, Ollama, Unsloth, llama.cpp, MLX, Kaggle, and more.
|
|
📌 A 16 GB MacBook or consumer GPU such as an RTX 4060 is enough to run it locally.
|
|
⚡ Encoder free design means you manage one unified model for text, image, and audio instead of juggling a separate vision tower.
|
|
🔭 Apache 2.0 licensing plus first class support in tools such as Ollama and llama.cpp makes this a drop in upgrade for many local stacks.
|
|
|
Official blog →
Get weights on Hugging Face →
|
|
|
🎙️ MISO ONE · OPEN SOURCE TTS · VOICE CLONE
|
|
Free 8B TTS Clones Your Voice in 10 Seconds With Real Emotion.
|
|
Miso One from Miso Labs is an 8B open source text to speech model released completely free. With 10 seconds of your voice it can clone you with natural prosody and emotional range including joy, anger, and sarcasm. Latency is around 0.11 seconds, which makes it ready for real time conversation. It can run locally or through a forthcoming API. Demos of YouTube narration, podcast voices, education content, and live agents show it matching or beating many paid services at zero software cost.
|
|
📌 Ten second voice clone, roughly 0.11 second latency, and open weights that you can run on your own hardware.
|
|
⚡ Strong fit for automated channels, podcast production, real time agents, and personal voice products for creators and educators.
|
|
🔭 Paid tools such as ElevenLabs charge monthly for similar quality. Miso One puts that level of TTS inside a free and self hostable model.
|
|
|
Try Miso One free →
|
|
Learn AI in 5 minutes a day
You don't have to scroll every AI thread, track every new tool, or watch every demo.
The Rundown AI breaks it all down for you — the latest AI news, tools, and tutorials in one free 5-minute email every morning.
Trusted by 2M+ professionals at Apple, Google, and NASA.
|
Today's Top Launches
Products To Watch Out Today
|
🏆 Embedded iPaaS from Albato
API · SaaS · No‑Code
Drop a full integration layer into your SaaS so users connect their tools inside your product instead of relying on Zapier style hacks.
|
|
💺 DeskHog
Office · Workplace · Tools
Hybrid office seat planning without spreadsheets. Lets teams book desks, see who is in, and avoid the usual hot desk chaos.
|
|
💡 Ideabrowser.com
Research · Productivity · Artificial Intelligence
A focused research browser for exploring ideas with AI help. Good for founders who want deep dives without getting lost in normal tabs.
|
|
Stay sharp,
Builder Hustle
