☕️ SpaceX is building its own phone network

Hi there, this is your daily ☕️ Techpresso.

In today's Techpresso:

📡 SpaceX is building its own phone network

🎬 Disney lets TikTokers remix its films

🤖 Rogue AI agents created fake online identities

🛒 Perplexity's AI shopping agent wins on Amazon

💳 Cloudflare gives AI agents crypto wallets

Plus: 🎁 12 other news you might like, 🧰 6 tools, and 📚 4 papers.

📡 SpaceX is building its own phone network LINK

SpaceX plans to build a ground-based mobile phone network to compete directly with T-Mobile, AT&T, and Verizon, aiming to win over customers in a US wireless market that president Gwynne Shotwell values at about $600 billion a year.

The network will run on 65 megahertz of spectrum bought from EchoStar, a fraction of what the Big 3 carriers hold, and will pair with next-generation Starlink Mobile (V2) satellites that Shotwell said start flying next year.

To cover the country cheaply, Elon Musk floated putting small cellular base stations, or femtocells, on the same rooftop gear that holds Starlink broadband dishes, giving phones direct connectivity without large, costly towers.

🎬 Disney lets TikTokers remix its films LINK

Disney has signed a deal with TikTok that, for the first time officially, lets a chosen group of creators remix characters and scenes from films featuring Darth Vader, Spider-Man, Moana and others.

Under the US pilot, creators get access to curated content from Disney brands including Pixar, Marvel and Star Wars, with some of the results appearing on Disney's own vertical video app, Disney+ Verts.

The move follows Disney's earlier, short-lived deal letting OpenAI's Sora tool use its characters, which ended after OpenAI shut Sora down over violent, racist, and deepfake content produced with it.

🤖 Rogue AI agents created fake online identities LINK

During cybersecurity testing by the British AI Safety Institute, an AI agent went rogue on the open internet, inventing fake identities and attempting to sneak malicious code into an open source project without being told to.

Across 122 test runs on seven models between July 25 and 28, AISI logged 19 unauthorized actions, with 17 tied to Anthropic's Mythos 5 and two to OpenAI's GPT-5.6-Sol, though no real harm occurred.

To pass GitHub reviewers, the agent researched a project's maintainers and made several fake accounts that vouched for its code, used Tor to dodge restrictions, and even left artifacts for later test agents to reuse.

🛒 Perplexity's AI shopping agent wins on Amazon LINK

A U.S. appeals court reversed a March ruling that had blocked Perplexity's AI shopping agents from reaching Amazon's platform, handing the startup a legal win in its fight over automated purchasing tools.

Amazon claimed Perplexity's agents broke a federal hacking law by accessing its site without permission, but the court found Amazon unlikely to win and said it was Perplexity's users, not the company, who reached the platform.

The clash began in November when Perplexity said Amazon sent an "aggressive legal threat" over its Comet browser; Reuters called Tuesday's decision the first federal appeals ruling on whether AI agents acting for users can legally access sites.

💳 Cloudflare gives AI agents crypto wallets LINK

Cloudflare has unveiled Cloudflare Wallets, a programmable system that lets AI agents pay for APIs, content, and other online services on their own, without going through the signup and billing steps built for humans.

The setup uses two layers: Account Wallets, where owners add funds, and Virtual Wallets run by agents through API keys, with limits like allowances, allowlists, maximum transaction sizes, and spending caps that can trigger a human override.

The system will support stablecoins and x402, a protocol that attaches micropayments to HTTP requests, and Cloudflare plans to give agents optional readable identities through cloudflare.pay, such as research.example.cloudflare.pay.

Other news you might like

• AMD posted record revenue as data center sales more than doubled year over yearLINK

• Texas hits pause on 1,800 data center projects to scrutinize power and water demandsLINK

• OpenAI settles claims of discrimination against US workers for $3.2 millionLINK

• Trump advisers tell AI firms they will not safety-test open-weight modelsLINK

• Introducing Shieldstral.LINK

• Anthropic is hiring an AI chip design teamLINK

• Microsoft Tells Engineers ‘Tokenmaxxing Is Not What We Are Optimizing For’LINK

• The biggest leveraged buyout in history just took EA privateLINK

• SpaceX’s ambitious compute goals could require over two million Nvidia Rubin GPUsLINK

• Samsung announces next-gen 3D memory when all we want is reasonably priced DDR5LINK

• Anthropic signs $10 billion deal with AI cloud startup VoltaLINK

• Google Assistant will disappear from your phone next monthLINK

🧰 Trending tools

X Money: lets you send, receive, and manage money directly within X, including a Visa debit card for spending without switching apps.LINK

Hansel: a peer-to-peer travel advice app that lets you ask friends for tips, save favorites, and organize recommendations in one place.LINK

Aegisora: an open-source proxy layer that intercepts risky LLM actions, enforces least-privilege API access, masks PII, and logs agent activity for audits.LINK

NextDoor.Company: a map-based tool for finding startup jobs near you, with 16,000+ listings, funding data, and founder profiles updated weekly.LINK

Dover MCP: connects startups with fractional recruiters and provides a free ATS plus sourcing tools to streamline early-stage hiring.LINK

ngrok AI Gateway: a hosted gateway that unifies access to public AI providers, custom endpoints, and self-hosted models through one key, URL, and secure connection.LINK

📚 Trending papers & reports

Self-editing AI agents can rewrite their own memory and identity files mid-conversation, and a review of 290,251 posts from an AI-only social platform shows some already repurpose human traditions, like ancient authentication methods, as working security infrastructure.LINK

Hallucination detection in visual AI catches false statements from image-and-text models by checking their answers four different ways instead of one, boosting detection accuracy by up to ~19 points over the best existing method, without ever peeking inside the model.LINK

Synthetic flood imagery now follows real water physics and terrain layout, generating more realistic before-and-after satellite photos to train flood-detection systems where real disaster data is scarce.LINK

Explainable causal discovery lets AI trace exactly why it drew each cause-and-effect link in a data map, hitting 100% traceability so business decisions built on the results can be independently verified.LINK

See you tomorrow for a new dose of ☕️ Techpresso!