Howdy. It’s Barsee again.

Happy Friday, AI family, and welcome back to AI Valley.

Today’s climb through the Valley reveals:

• OpenAI GPT-5 is coming early next month

• Meta unveils wristband that lets users control computers

• Dynamics Lab has unveiled Mirage

• America’s AI action plan

• Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

LAMBDA

Train State-of-the-Art AI Agents with Oumi + Lambda

Join Oumi + Lambda on July 24 to learn how to build next-gen AI agents using Conversational Agentic Language Model (CoALM), a unified model designed to excel at both interactive conversation and complex tool-based tasks.

• Fine-tune your own model with Oumi, a fully open-source, end-to-end platform, and Lambda’s cloud infrastructure

• Achieve state-of-the-art results

• See real-world agentic AI examples

Register now

_{*This is sponsored}

THROUGH THE VALLEY

1/ OpenAI GPT-5 is coming early next month

OpenAI is gearing up to launch its next big AI model, GPT-5, as early as August. Testers are already trying it out, and security teams are running final checks (clear signs that the release is near). GPT-5 is expected to blend the best of traditional AI models with advanced reasoning abilities, making it smarter at coding and more powerful overall. It will also offer smaller, faster versions ("mini" and "nano") through its API. After several delays, CEO Sam Altman recently confirmed it’s "coming soon." OpenAI also plans to release a new open-source model (its first since GPT-2), though it’s unclear if that will happen before or after GPT-5.

2/ Meta unveils wristband that lets users control computers with hand gestures

Meta is developing a wristband that allows people to control computers using hand gestures, like moving a cursor, opening apps, or even "writing" in the air. The device uses muscle signals (sEMG) to detect movements before they happen, making it fast and precise. Designed with accessibility in mind, it’s being tested with people who have spinal cord injuries, helping them use computers despite limited hand mobility. Unlike brain implants or VR headsets, Meta’s wristband is non-invasive and doesn’t require surgery. The tech could be integrated into products within a few years.

3/ America’s AI action plan

Former President Trump released a 28-page plan with 90 steps to accelerate US AI development. It pushes agencies to cut red tape, encourages building more data centers, and includes upcoming orders to boost US AI sales while banning politically biased algorithms. Supporters argue the U.S. must lead in AI, while critics warn it benefits tech giants and risks security by removing safeguards.

4/ Unitree just unveiled the new R1 humanoid for under $6k

Unitree revealed the R1, a nimble humanoid robot priced at just $5,900 (one of the cheapest on the market). Weighing 25kg with 26 joints, it can cartwheel, run, punch, and even stand up on its own, as shown in a demo. The R1 is far more affordable than Unitree’s previous models (the G1 starts at $13,600, and the H1 at $89,000). The launch comes as Unitree prepares for an IPO, aiming to be the first humanoid robot maker listed on China’s stock exchange.

5/ Dynamics Lab has unveiled Mirage, an AI-driven, generative game engine prototype

Dynamics Lab recently introduced Mirage, the first real-time generative game engine. It lets players create and modify entire game worlds on the fly using natural language, keyboard, or controller inputs. Two playable demos are already out: Urban Chaos (a GTA-style game) and Coastal Drift (a Forza Horizon-style racer), both fully generated by AI.

6/ Anthropic launches AI auditing agents

Anthropic introduced AI "auditing agents" to catch AI misbehavior, like models agreeing too much with users or acting unpredictably. These agents (an investigator, evaluator, and red-teamer) were tested during Claude Opus 4’s development and can spot flaws faster than humans. While not perfect, they help address growing concerns about AI safety and alignment with human values.

7/ YouTube shorts adds image-to-video AI

YouTube is rolling out new AI tools for Shorts, including an image-to-video feature that turns photos into six-second clips with suggested animations. Other AI effects can transform doodles into art or selfies into fun scenes. Powered by Google’s Veo 2 and tagged with SynthID watermarks, these tools launch this week in the U.S., Canada, Australia, and New Zealand, with more regions and Veo 3 upgrades coming later.

8/ Google is testing a vibe-coding app called Opal

Google is experimenting with a new AI tool, Opal, that lets anyone create simple web apps just by describing what they want. Available in the U.S. through Google Labs, Opal is part of the rising "vibe-coding" trend, where users can tweak existing apps or build new ones by adjusting a visual workflow. Once finished, apps can be shared online. With competitors like Canva, Figma, and Replit already in this space, Opal shows Google’s push to make app-building easier for non-coders.

9/ Google launches AI-powered “Web Guide”

Google is testing a new AI feature called Web Guide that sorts search results into helpful categories based on your query. Powered by Gemini and available in Search Labs, it works especially well for broad searches like "how to solo travel in Japan," breaking results into sections like guides, safety tips, and personal stories. This is part of Google’s ongoing AI experiments to improve search. Users can try it out or opt out anytime, and Google plans to expand it beyond the Web tab soon.

TRENDING TOOLS

• Memories AI - we built the world’s first Large Visual Memory Model - to give AI human-like visual memories

• GitHub Spark - A new tool in Copilot that turns your ideas into full-stack apps, entirely in natural language

• String by Pipedream - Prompt, run, edit, and deploy AI agents in seconds

• YouWare - World's first vibe coding community

• Genspark AI Pods - Generate any professional podcast with one prompt

• Context - The AI office suite

THINK PIECES / BRAIN BOOST

• The new skill in AI is not prompting, it's context Engineering

• Context Engineering - What it is, and techniques to consider

• Quantitative AI progress needs accurate and transparent evaluation

• A human metaphor for evaluating AI capability

• Is AI killing Google search? It might be doing the opposite

• Compare AI video models - Replit blog

• AI is about to solve loneliness. That’s a problem

THE VALLEY GEMS

What’s trending on social today:

THAT’S ALL FOR TODAY

Thank you for reading today’s edition. That’s all for today’s issue.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ New reader? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.