• Apple Sues OpenAI
  • An Inkling of Models to Come
  • One Minus in the US Market

Apple Sues OpenAI
I would not say to anyone casually that Apple is litigious, but they are very willing to go after anyone who might dare try to expose internal information on their design, product, and marketing strategies. Apple does not like spoilers, and when it can, it goes after those who would spill the beans early like Jon Prosser, a prolific leaker with the popular YouTube channel Front Page Tech where he often reveals new products and software design before they are unveiled at official Apple events. And should someone get their hands on an actual prototype and dare show it to the world, well, terrible things tend to happen; just ask Gizmodo about the iPhone 4.

Historically, most of the attempts to reappropriate Apple design and aesthetic have come from outside the United States, so Apple has not gone after as many companies as perhaps it would like. But last Friday, not long after I'd published last week's TTNN, the news dropped that Apple was suing OpenAI for stealing trade secrets in an attempt to further its own efforts to establish a hardware line. And I have to admit that at first this didn't immediately make sense to me. Although many a billionaire has a dream of creating their own phone to get around the complexities of operating at a disadvantage on the iPhone, they all know they have to find a way to make it work as best as they can. History has shown that having a better platform doesn't matter because Apple and Google have such a mighty lead that it is nearly possible to overcome it, even with great sums of money and all the will in the world. Something would have to happen to turn the public away from these companies in order for their platforms to become vulnerable to real competition. Short of accomplishing that, one must make apps and services that run on the iPhone and Android. And it seemed that until very recently, OpenAI has had a privileged position on iOS with ChatGPT as the preferred backbone for AI queries beyond Siri's capabilities. ChatGPT was already the best known chatbot in the United States and probably everywhere else and having it built-in on one of the largest mobile platforms was a big win for the company, even if Apple peskily kept saying they were open to having other models available.

Even knowing that Apple has advanced its AI efforts significantly with the new Siri AI, built with help from Google's Gemini, it didn't seem obvious that Apple would have any products OpenAI would want to steal, despite the common knowledge that OpenAI has been working on hardware for quite some time and even acquired former Apple design executive Jony Ive's hardware and design studio Love From to inject new hope into its hardware ambitions. Apple has an AR/VR headset in the Vision Pro and we believe them to be working on smart glasses and they already have screenless smart speakers and an operating system that could be used to show information on a smart speaker with a display, but we haven't heard much of any other ideas for new AI gadgets that might be under consideration in Cupertino.

Yet the allegations are a little bit crazy: supposedly, a former Apple employee now working at OpenAI lured away some other Apple employees and got them to bring confidential material with them, which in some cases, was used in "show and tell" sessions in interviews before the employees were even hired. Another former employee was accused of exploiting a software bug on a company-issued laptop to download a wide variety of internal files on Apple's hardware plans that he then also brought with him to OpenAI.

Honestly, this all seems ridiculous. Even if Apple had some super top-secret AI device in the works that OpenAI wanted to use as a basis for its upcoming hardware, the truth always finds you out, and more importantly, Apple's success outside of traditional machines like phones, tablets, and laptops has been moderate at best. There is no guarantee that any smart device Apple has would be an instant hit and certainly if someone else launched something based off that gadget, profit would not be an automatic outcome. It has been largely speculated that OpenAI is working on some kind of wearable, perhaps something like the Plaud Note Pin, Amazon's Bee, or something with aspects of multiple existing gadgets. Products like these haven't had the biggest impact on consumer electronics; certainly, Plaud has had some success with its tightly focused Note line and Bee was considered interesting enough for Amazon to buy it, but these small AI gadgets aren't in demand and don'ts seem to be able to offer the kinds of experiences that would make them be in demand. For people to carry around yet another gadget, it has to offer value that isn't easily attainable from the phone or can fit into devices and device types people already use, like rings and glasses. A whole new device is unproven and risky. Instead of adding more risk to the process, OpenAI should have just gone the usual route of looking for holes in the market, gathering feedback from real people about ways to plug those holes, and then going to market with a strategy built around those findings.

Why does it seem like no one does this anymore?

An Inkling of Models to Come
Large language models have been steadily growing in both popularity and in engendering dystopian fears for a few years now, but they have their limitations. More people in the industry are realizing that to achieve artificial general intelligence (AGI), language is insufficient alone. AI will need awareness and knowledge of the physical world to catapult robotics over the next technological hurdle. But you can't just take your chatbot outside and have it walk around, so as AI companies iterate, they are beginning to incorporate more additional data types into their training processes. Text and images are joined by video and audio for a more complete understanding of our world. Some models are beginning to incorporate map and geographical data as well. As more data types are ingested, new and better AI models emerge. Mira Murati, the former CTO at OpenAI during the Sam Altman Ouster Era (as I shall now call it from now on), started her own company Thinking Machines and they have recently unveiled their newest multi-modal model, Inkling.

The model is aimed at corporate customers, and it has been trained to deal with difficult and sensitive topics without requiring corporations to apply their own filters. It is designed for agentic workflows and coding assistance and has safeguards designed to reduce hallucinations and simple guessing at answers. And although it has internalized images, video, and audio, its output is currently limited to text. It can support a context window with up to 1 million tokens and the model itself has 975 billion total parameters (for some context, DeepSeek's latest model has roughly 1.6 trillion). As a nod to the expense issues companies face using the latest AI frontier models, its uses only 41 billion of its parameters per token and users can further refine the experience by adjust the model from minimum to maximum reasoning settings. Inkling is also an open-weight model, which means its training parameters have been publicly released, and users are able to fine-tune with their own data as needed. The most popular frontier models are closed weight, and their training data is proprietary. Thinking Machines even has their own training platform called Tinker.

Whether or not Inkling and ultimately Thinking Machines as a company is successful, this effort shows ways that AI companies can innovate and pivot to better support their customers. Multi-modal models that incorporate more kinds of data and models that are geared for more integrity and efficiency will continue to be important in the enterprise market as companies re-evaluate their technology budgets and AI expenses and look for ways to reduce cost and waste. Thinking Machines is offering a compelling alternative to popular open-weight Chinese models like DeepSeek or Alibaba's Qwen and as an American company, that may be enough to push them forward in the market for now.

One Minus in the US Market
I have written before about the potential demise of the OnePlus brand in the United States, but now the official word has come, and the brand is indeed making its exit.

This won't be the usual I-will-miss-them type of post because in all honesty, I never bought a single OnePlus product (except for the OnePlus Buds) so I can't say that I specifically will miss them. But I will miss their presence in the United States because one fewer competitor makes all the others a little bit lazier. But we're missing far more than just one manufacturer. We are missing out on so many brands and so many features that one might even say we live in a mobile desert. Much like food deserts, no one wants to risk building anything meaningful here. And it's not only our contradictory and cantankerous government that leads to this outcome, but the nature of how we buy and sell our mobile devices. Because we are nation of debtors, it makes sense on some level that we would be the country to go all-in on debt financing for the purchase of a phone much in the same way one might buy a car or home. The biggest difference is that one doesn't usually pay the same interest rates, but the telecoms find other less obvious and more devious ways to make us pay extra for the privilege of spreading out the pain of buying a new phone over 2-3 years. With more fees and fewer features, our monthly contracts cost us more money, and we find it harder to move from carrier to carrier to try to reduce monthly expenses because of early termination fees or the need to pay off the phone in full to leave. Most of the rest of the world doesn't function this way. You buy a phone, you take it to a carrier, you pay for the service and if you don't like the service, you take your phone elsewhere. This means that companies can sell their phones in electronics stores or even direct to consumer and people will find them. In the United States, if your phone isn't sitting in a Verizon, AT&T, T-Mobile, Walmart, or Best Buy store display, it might as well not even be for sale in the US. Truth be told, even an unlocked phone at Best Buy isn't likely to sell unless Best Buy can get you to bundle it with service from one the providers they offer in-store. Brands like OnePlus and now Nothing were/are surely glad to be available more broadly here, but they also know the sales bump is likely to be very small.

To up-end this, someone with the budget and scale of an Apple, Samsung, or Google would have to flip this model on its head. They would have to take on the financing (or offer financing through partners like Klarna, Affirm, or PayPal) and offer phone plans as an add-on. Yes, Best Buy kinda-sorta does this now, but it would need a change in strategy and real market muscle behind it to do this on a serious and meaningful level. Yes, Apple, Samsung and Google do have some stores in the US as well where you can buy their phones, but none is really pushing phone plans. We would need a store where you go to buy the best phone for you and then pair it with the best plan for you. If for any reason you are unhappy with the service, the store can help you switch to another carrier, but you don't lose anything with the phone because your mobile plan was always separate.

This would never happen, I know, but it would open our market to more players and allow them to sell their unlocked phones here and get them in the faces of more people who would buy them if they were simply aware of their existence. This would encourage more manufacturers to make phones that work with our 4G/5G bands and give us some real choice for the first time. It might have even kept OnePlus in our market for a little bit longer.

As it is, we have to hope Nothing and Motorola (Lenovo) can keep hanging on here.