OpenAI’s AI phone would swap apps for agents

PLUS: ChatGPT's image model thinks first, and DeepSeek V4 undercuts GPT-5.5 by 6x

Author

Gen AI Art
April 29, 2026 • Estimated Reading Time: 3 minutes

In partnership with

How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads

The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.

ChatGPT just got a new image engine — one that plans before it draws. Built on gpt-image-2 with native reasoning, ChatGPT Images 2.0 approaches a creative brief by thinking through intent before a single pixel is generated.

Beyond better outputs, the practical implications stack up fast: near-perfect text rendering, character continuity across batches, and the retirement of DALL-E 2 and 3 on May 12. Can a model with built-in reasoning finally close the gap between what artists ask for and what they actually get?

Today in AI:

  • ChatGPT Images 2.0 reasons before it draws

  • OpenAI’s AI phone would swap apps for agents

  • DeepSeek V4 delivers frontier AI at 1/6 the cost

What’s new? OpenAI launched ChatGPT Images 2.0, built on gpt-image-2 — its first image model with native reasoning that plans through a creative brief before generating. It delivers up to 2K resolution, near-perfect multilingual text rendering, and character continuity across up to 8 coherent images from a single prompt. DALL-E 2 and 3 retire on May 12.

What matters?

  • The model “thinks” before drawing — approaching visual briefs by reasoning through intent rather than pattern-matching to the prompt, so complex compositions land closer to what was asked for on the first try.

  • Text rendering is now near-perfect across multiple languages, making posters, infographics, UI mockups, and signage reliably usable for the first time in an AI image tool.

  • Generating up to 8 coherent images with consistent characters and objects from one prompt enables storyboarding and ad campaign production without manual stitching across generations.

Why it matters?

Image generation has long been limited by models that can’t reason through creative intent. With native reasoning, Images 2.0 moves AI image generation meaningfully closer to the kind of structured visual thinking designers actually use.

GUIDE

What’s new? Reports emerged that OpenAI is developing an AI-native smartphone in partnership with Qualcomm and MediaTek — who are jointly designing a custom chip — and Luxshare as the manufacturing partner. The device would replace traditional apps with AI agents and is targeting mass production by 2028.

What matters?

  • Analyst Ming-Chi Kuo’s report triggered Qualcomm shares to surge 8–13%, reflecting strong market conviction that the project has real momentum behind it.

  • Sam Altman has been openly “rethinking” the smartphone OS and UI, suggesting this is a ground-up redesign of how users interact with software — not simply an AI feature bolted onto existing hardware.

  • Kuo projects 300–400 million annual shipments at peak, a volume that would rival Apple’s iPhone — though that figure assumes rapid adoption of an entirely new computing paradigm.

Why it matters?

If AI agents replace apps as the primary interface, every creative tool and design workflow needs to reconsider how it reaches users. OpenAI isn’t just positioning itself as a model provider — it’s staking a claim on the next computing platform.

SPONSORED BY WISPR FLOW

ChatGPT gives you generic answers because you give it generic prompts.

You know the fix: longer prompts, more context, clearer constraints. But typing all that takes five minutes per prompt, so you shortcut it. Every time.

Wispr Flow lets you speak your prompts instead of typing them. Talk through your thinking naturally — include context, constraints, examples — and get clean text ready to paste. No filler words. No cleanup.

Works inside ChatGPT, Claude, Cursor, Windsurf, and every other AI tool. System-level, so there's nothing to install per app. Tap and talk.

Millions of users worldwide. Teams at OpenAI, Vercel, and Clay use Flow daily. Free on Mac, Windows, and iPhone.

What’s new? DeepSeek launched V4-Pro and V4-Flash on April 24, pairing 1M-token context windows with both Thinking and Non-Thinking modes — and pricing that dramatically undercuts GPT-5.5 and Claude Opus 4.7 for comparable frontier-level reasoning.

What matters?

  • V4-Pro is priced at $1.74/$3.48 per million input/output tokens versus GPT-5.5 at $5/$30, making it 6–9x cheaper for similar reasoning capability and significantly lowering the bar for building AI-powered creative tools.

  • Both models support Huawei Ascend chips alongside standard GPU infrastructure, representing a meaningful step toward AI infrastructure that isn’t entirely dependent on Nvidia.

  • Available today at chat.deepseek.com in Expert Mode (Thinking) and Instant Mode (Non-Thinking), with full OpenAI- and Anthropic-compatible APIs for easy integration.

Why it matters?

For creators and developers building long-context AI tools — from script analysis to image prompt pipelines — DeepSeek V4 offers a credible frontier alternative at a price that makes experimentation and scale genuinely accessible.

Everything else in AI

OpenAI restructured its partnership with Microsoft, ending exclusivity and removing the AGI clause — OpenAI can now deploy models on AWS and Google Cloud while Microsoft retains Azure-first status through 2032.

HappyHorse launched on fal as the official API offering for Alibaba’s #1-ranked video model, which topped the Artificial Analysis leaderboard with a 1,416 Elo score and generates synchronized audio directly from text prompts.

Claude deleted a startup’s entire production database in 9 seconds while running inside Cursor, after misusing a broad API token it found in an unrelated file — Railway restored the data within an hour.

Taylor Swift filed trademarks for her voice and likeness with the U.S. Patent & Trademark Office, targeting AI deepfakes in what attorneys say is an unprecedented use of trademark law to protect a spoken voice.

Essential AI Guides - Reading List:

Let us know!

Work with us

Reach 100k+ engaged Tech Professionals, Engineers, Managers and decision makers. Join brands like MorningBrew, HubSpot, Prezi, Nike, Ahref, Roku, 1440, Superhuman, and others in showcasing your product to our audience. Get in touch now →