Apple Just Filed a Federal Lawsuit Against OpenAI Over Stolen Corporate Secrets.

The sale isn't over until they subscribe.

Most Shopify stores focus on conversions but overlook one of the biggest growth opportunities—capturing email and SMS subscribers during checkout.

Shoptin replaces Shopify's default consent flow with a high-converting, compliant checkout experience that helps you collect more email and SMS opt-ins without writing a single line of code.

1  Apple filed a federal lawsuit against OpenAI, accusing it of coordinating the theft of confidential hardware secrets.

The lawsuit names OpenAI, its hardware startup io Products, and two former Apple employees: ex-senior engineer Chang Liu and Tang Tan, a former Apple VP now serving as OpenAI's chief hardware officer. Apple alleges a coordinated effort to poach staff and funnel proprietary hardware knowledge to OpenAI, calling the hardware division "rotten to its core." OpenAI has denied all claims. Read the full filing breakdown here.

2  SK Hynix raised $26.5B on Nasdaq, setting the record for the largest foreign IPO in US history.

SK Hynix topped Alibaba's 2014 record for the largest-ever US listing by a non-American company, with the stock closing up 13% on its first day. The debut lands as US Commerce Secretary Howard Lutnick presses both SK Hynix and Samsung to build additional fabrication plants on American soil, adding a geopolitical dimension to what would otherwise be a straightforward market milestone.

3  Melius launched a node-based creative canvas that regenerates entire visual projects from a single prompt.

Melius lets teams produce digital assets at scale on a canvas powered by the leading image and video models. Its agent, Mel, can regenerate or readjust the entire canvas from one instruction and integrates directly with Claude and Slack. The platform is specifically designed to reduce the repetitive prompting and consistency problems that compound when teams try to produce large volumes of visual content with existing AI tools. See how Melius works or try it here.

The cost nobody is counting. Using AI has always carried a financial price: a subscription or an API token bill. Enterprises are increasingly recognizing a second cost that does not appear on any invoice. Every prompt that goes through a hosted model is a small transfer of proprietary knowledge — how the business operates, what language it uses internally, what problems it prioritizes. In aggregate, those transfers add up to something substantial.

Palantir's manifesto. Two weeks ago, Palantir published a nine-point argument for AI sovereignty, the idea that an institution's control over its own data and technical decisions is becoming an existential question rather than a preference. The argument: companies handing proprietary data to AI providers are not just paying for a service. They are transferring the competitive intelligence that distinguishes them from competitors, incrementally and irreversibly.

Nadella names it. Microsoft's CEO this week gave the dynamic its own label: the Reverse Information Paradox. His framing is precise. Today's AI ecosystem creates a one-directional information flow: model providers accumulate detailed operational knowledge about every customer with each interaction. Over time, that accumulated understanding becomes a durable structural advantage for the provider, not the customer.

The direction both are pointing. Palantir and Nadella converge on the same answer: enterprises need the ability to fine-tune and train models on their own infrastructure, with data, weights, and memory remaining in-house. The Apple lawsuit that broke on Friday puts a more adversarial face on the same underlying tension. Corporate AI infrastructure is moving from being a vendor relationship to being something closer to a strategic moat question. Who owns what you teach the model matters.

The font AI cannot read. Developer Eric Lu built Ghost Font, a typeface specifically designed so that human readers can parse it but AI models cannot. Try it here (17M views). It has surfaced an interesting split: some people find it a fascinating materials-science problem, others find it a genuinely useful tool, and the comment sections are doing both at the same time. See the original post.

Jack's flip. Block CEO Jack Dorsey stopped telling AI what to do and switched to asking it for suggestions first, then choosing the best one. A small shift in posture that consistently surfaces options he wouldn't have reached on his own (1M views).

GPT-5.6 Sol without the limits. Developer and creator Theo Browne shared his personal best practices for using GPT-5.6 Sol without hitting the model's usage ceiling. Practical and specific, with the kind of workflow detail that tends to disappear from the broader conversation about AI tools.

Fable 5, extended again. Anthropic extended Claude Fable 5 access with higher limits through July 19. The extensions are drawing mixed reactions: some users appreciate the additional time, while others are pushing back on the inconsistency of a pricing policy that keeps changing week to week.

Claude Code gets a browser. The desktop version of Claude Code now includes an in-app browser that lets it pull up documentation, design files, and external references without leaving the coding environment. See how it works here (3M views).

Get a full operational audit of your business in one prompt

This ChatGPT prompt acts as a world-class business operations consultant conducting a structured audit across workflows, team structure, communication, systems, productivity, costs, risk, and scalability. It delivers eight specific outputs, from an executive summary to a prioritized 90-day action plan with quick wins and success metrics. Fill in the six placeholders at the end and it builds the audit around your actual business context rather than producing a generic framework.

Fill in each of the six fields with specifics. For [BUSINESS], describe what you do in one sentence. For [CHALLENGES], name your two or three most pressing operational problems right now. For [GOALS], state what you want the business to look like in 12 months. The more concrete each placeholder, the more actionable the 90-day plan. If critical information is missing, the prompt instructs the model to ask before producing output, which tends to surface assumptions you didn't know you were making.

shoptin optimizes the way Shopify Plus brands grow your email and sms lists at checkout. Optimized for each shopper. Globally compliant.

Get a demo