☕️ Apple unveils first smart home display

Tuesday 13 October 2026 | Join Free | Upgrade

Together with

Hi there, this is your daily ☕️ Techpresso.

In today's newsletter:

🏠 Apple unveils first smart home display

🎮 Epic pivots to "AI first" game development model

🤖 Character.AI files for bankruptcy

🪱 Anthropic warns AI worms could hit homes

🧠 Meta launches two AI models

🇨🇳 Nvidia's China AI-chip share falls as Huawei closes gap, report finds

Plus: 🎁 11 more stories you might like, 🧰 6 tools, and 📚 5 papers.

🏠 Apple unveils first smart home display LINK

Apple has released the HomeView, its first smart home device with a screen, a 7-inch rounded display on a speaker base built mainly to talk to the overhauled, AI-powered Siri.
Priced at $299 and sold from 13 November in four colours, it uses a camera to recognise up to six people and show only their own messages and data, with a wall-mounted speaker-free version also offered.
Linked to an iPhone or iPad running Apple Intelligence, the HomeView can act across apps like Messages and Calendar, make FaceTime calls, work as an intercom, play Apple Music, and show photo-based clock faces.

🎮 Epic pivots to "AI first" game development model LINK

Epic Games is moving to an "AI first" development approach for its Disney project, letting AI models write all the code and guide the artwork for a planned shooter, according to a Bloomberg report citing people familiar with the work.
Human developers will manage and steer code and art made by models like Anthropic's Claude, meaning no more hand-written code and art assets built through full AI workflows, based on internal Epic messages Bloomberg reviewed.
The $1.5 billion Epic-Disney deal, announced in early 2024 for a Marvel- and Star Wars-themed shooter once set for March 2027, is now delayed indefinitely amid the AI shift and slow, understaffed progress so far.

🤖 Character.AI files for bankruptcy LINK

Character.AI has filed for Chapter 11 bankruptcy in a Delaware federal court, naming Disney as the first bidder for some of its assets as it tries to sell all or most of the company.
Disney agreed to pay $20 million for a permanent, non-exclusive licence to some of the chatbot platform's intellectual property and legal claims, and will lend back former staff it hired to keep the service running safely.
Owner Character Technologies listed assets and liabilities between $50 million and $100 million, expects the sale to close in late November or early December, and warned the platform will shut down if no buyer keeps it going.

🪱 Anthropic warns AI worms could hit homes LINK

Anthropic researchers warn that small AI models could soon spread themselves across poorly secured consumer devices like smart TVs and robot vacuums, acting as self-replicating "AI worms" that copy their own code and set themselves up on new hardware.
In tests, the researchers ran the Gemma4-26B-A4B model across 15 cheap Android TV boxes costing about $55 each, and found that one model, GLM-5.2, successfully completed the full replication chain on all eight test nodes in four out of five runs.
Anthropic CEO Dario Amodei warns that such a swarm could be ready within six to twelve months and might cause hundreds of billions of dollars in damage by taking over the internet through a persistent, decentralized botnet with no central kill switch.

🧠 Meta launches two AI models LINK

Meta released two new AI models on Monday, Muse Wonder and the larger Muse Wonder Full, saying they are its first built to rival the top models from Anthropic and OpenAI.
By Meta's own comparison table, Muse Wonder Full ranks just below Claude Opus 5.5 and above Claude Fable 5.1 on the Artificial Analysis Intelligence Index, but its model card admits its competitor runs were only "best-effort."
Muse Wonder Full needs a Muse Code subscription at $50 or $200 a month, while Muse Wonder costs about $4 per million input tokens, and Meta says both will reach the Muse app soon with no set date.

🇨🇳 Nvidia's China AI-chip share falls as Huawei closes gap, report finds LINK

Nvidia's hold on China's data-centre AI chip market dropped to 34% in the first half of 2026, as US export rules choked supplies and Huawei and other local makers moved in, according to an IDC report seen by Reuters.
Nvidia shipped 678,000 AI accelerators in China from January through June, a 44% drop from a year earlier, while Huawei nearly doubled its shipments to 572,000 units, reaching a 28% share of the overall market.
Smaller Chinese players also grew, with Alibaba's T-Head at 7%, Cambricon 5%, Baidu's Kunlunxin 4%, and Hygon 3%; Jensen Huang has blamed Washington's curbs for handing the $50 billion market to Huawei.

Other news & articles you might like

Apple trails Samsung as global smartphone shipments hit a 13-year low LINK

Alibaba starts selling its Qwen N1 AI glasses in China from $450 LINK

AI labs sought control over what outside evaluators publish, Reuters says LINK

SpaceXSI bundles X Premium, Grok, and Cursor into one subscription LINK

Intel, Broadcom CEOs warn capital crunch could become AI's biggest bottleneck LINK

IMF warns $1.4tn in circular AI deals could deepen a market correction LINK

Google invests in period tracker Flo and plans women’s health wearables LINK

'123456' password linked to Denmark data breach affecting 8.8M people LINK

Ledger says hardware implant in cryptocurrency wallets stole recovery phrases LINK

Apple launches a new Apple TV 4K with Siri AI and remote-free voice commands LINK

Anthropic says it cut Claude reasoning token theft by up to 99% in fight against industrial-scale distillation LINK

🧰 Trending tools

Career Agents: you already have a job. Finding the next one is a second one, so a dedicated agent in your field handles applications, outreach and interview prep for you. Book a free consultation

Music Dawn: a recorder for musicians that captures ideas in one tap, applies finished sounds instantly, and syncs every take across iPhone, iPad, and Apple Watch. LINK

Kitbar: a Mac and Windows status bar showing Vercel and Cloudflare deploys, GitHub and GitLab checks, Stripe payments, and AI usage limits at a glance. LINK

Agent Craft: generates personal brand content for social media at scale, matching your voice to boost engagement and capture leads automatically. LINK

GotEmail: unifies Gmail, Outlook, iCloud and IMAP inboxes in one native macOS window with quick sign-in codes, instant alerts, and on-demand AI drafting. LINK

The Drive AI: an AI file assistant that organizes, searches, and creates documents across Google Drive, OneDrive, SharePoint, and Dropbox using plain English commands. LINK

PostNitro for Agents: equips AI agents with a design engine to generate on-brand carousels, images and video, then schedule posts across major social platforms via MCP, CLI or REST. LINK

You can check the previous tools here, or add your tool here

📚 Trending papers & reports

> Reach 900,000+ tech professionals: If your company is interested in reaching an audience of tech executives, decision-makers and engineers, you may want to advertise with us.

> Faster Transformer inference speeds up large language models by singling out the few oversized data signals that cause most accuracy loss, cutting one model's approximation error by ~9% versus a leading rival method. LINK

> Robot training from human videos teaches machines manipulation skills by watching ordinary first-person footage instead of costly robot data, lifting real-world task success from ~63% to ~85% as video hours scale from 1,000 to 5,000. LINK

> AI agent training teaches software assistants that complete multi-step tasks which early decisions actually led to success, lifting task completion rates by up to ~5 points on shopping and household simulations without extra data collection. LINK

> Compressed AI models can run on faster 4-bit chips without needing sample data to tune them, matching or beating methods that do require it, so companies cut inference costs with less setup work. LINK

> Faster AI serving speeds up a popular class of efficient language models by up to ~1.5x on standard replies and ~1.8x on the largest systems, cutting the memory bottleneck that slows them down. LINK

🤝 From our community: Building a macOS infrared RAW developer

John used AI to build a macOS app that develops RAW files from infrared-converted cameras, handling the specialized color processing those images need and cutting the time he spends editing shots by hand.

Read the full use case →

You can see more community use cases here, or submit your own here.

🎓 Want to master the AI tools we cover every day?

Techpresso's AI Academy has 330+ step-by-step tutorials on ChatGPT, Claude, Perplexity, and every tool that matters. No fluff — just practical workflows you can use at work. Try it free for 7 days.

On this day in 1994, released its first browser, Mosaic Netscape 0.9.