OpenAI is Building the iPhone of AI

The evolution of OpenAI’s suite of models follows a trajectory strikingly similar to that of Apple’s iPhones. With every successive and advanced update, the consumers are willing to pay an extra price. “The o1 preview is deeply flawed , but when o1 comes out, it will feel like a major leap forward,” said OpenAI’s chief, Sam Altman, on his long-term vision of providing quality over quantity.?

Altman himself draws heavily from Steve Jobs’ legacy and is actively building the ‘iPhone of artificial intelligence’. His focus on shipping “rockets ” instead of trash isn’t just a throwaway observation about the current AI market but a core part of OpenAI’s new strategy.?

“Our first product, GPT-3, was not something we were particularly proud of at the time,” admitted Altman in an interview . The transition from GPT-3.5 to GPT-4, and the upcoming GPT-5, could be an interesting case study of how every new update renders the earlier versions less significant – a trend particularly evident in the case of iPhones .?

Older Models Turn (Almost) Redundant

Much like the rapid and unforeseen rise of smartphones—now so essential we can hardly imagine life without them—AI is moving along a similar path. With companies like OpenAI rolling out ever-evolving models, their widespread adoption feels inevitable. Besides, they’re making older versions redundant.?

Roman Huet, the head of developer experience at OpenAI, confirmed on X that the upcoming DevDay would not see any new models or ChatGPT features, but only API updates, unlike last year’s event, which saw a series of big announcements. The evolution of ChatGPT, and OpenAI’s models since 2022, is very telling of the company’s direction towards shipping big.?

Conversely, it would be interesting to see that even for its incremental advanced features, if the company charges for them, there would be buyers, as the previous free versions wouldn’t provide the same value.

“I think the one big difference between smartphones and AI models is that the former is very incremental, but what is happening with AI is that the next feature is very exponential—in terms of capabilities and adoption. We have seen this with our own models—the capabilities of the newer ones are much better than the previous ones,” said Manu Kumar Jain, CEO at G42.?

Jain previously spearheaded the smartphone adoption in the country during his tenure at Xiaomi, and is now following a similar strategy with their largest Hinglish AI model, NANDA (named after the second-highest peak in India, Nanda Devi).?

Manu Kumar Jain, G42 India CEO, in conversation with AIM editor Amit Raja Naik at Cypher 2024?        

“With NANDA, we are creating a model that speaks to India–its languages, its people. This is not just about AI for a few; it’s AI for a billion,” said Jain, adding that the adoption of AI in real-life cases is sometimes higher than smartphones. “In the coming years, we cannot imagine our lives without AI, and I believe its adoption curve will be even higher than the smartphones,” he added.??

ChatGPT brought in this revolution, and with every successive update, we are seeing the older models get obsolete. In terms of adoption, catering to user experience will be the key differentiator for AI companies moving forward, and OpenAI is already incentivising in this area.?

Altman has always been obsessed with clean aesthetics . The simple and intuitive design language of OpenAI, consistent across its suite of products, especially ChatGPT, adds to its appeal.?

Taking this up a notch on the minimal route, OpenAI’s potential logo rebranding from its current symmetrical hexagon to a bold, simple black “O” is a bid to convey an impending brand and ideology shift within the company. It is interesting to note that OpenAI’s typefaces are currently licensed.?

Former iPhone designer Jony Ive joining OpenAI sort of solidifies this new creative journey that the company has embarked upon. While the project he would work on is undisclosed, his influence on the company’s future design would be huge.?

All of this appears to be an organic push for Altman towards a paradigm where he is slowly adopting a strategy of offering premium experiences to his consumers, which is quintessential of the early iPhone wave. “We’ve started to deliver great products, but we’re going to deliver far greater ones,” Altman said .?

Altman has always been an iPhone fanboy. OpenAI’s offerings, like o1, Sora, DALL·E, and the new ChatGPT advanced voice feature, are proof that advanced intelligence and accessibility can be availed for as low as $20 (for now, as higher subscription prices may be announced at OpenAI’s DevDay 2024).?        

And just like the iPhone, consumers would be willing to pay for it.

Enjoy the full story here .?

AIM 100: The Most Influential Global Leaders in AI

AIM proudly brings to you AIM 100 – the list of most influential global leaders in AI, who are driving innovation, tackling ethical challenges, and shaping the future of artificial intelligence with a focus on responsibility, inclusivity, and technological advancement, ensuring AI's development benefits all of humanity.

Access list here .?

AI Bytes

• Ahead of DevDay 2024, OpenAI has launched the Whisper V3 Turbo , a transcription model that is eight times faster and half the size of its predecessor, providing faster and more efficient transcription with minimal loss in accuracy, making it accessible for diverse applications while maintaining support for over 99 languages.
• Tredence has unveiled its AI Experience and Collaboration Center in Bengaluru, designed to drive co-innovation by enabling clients and partners to engage with agentic AI solutions, explore rapid prototyping, and co-create AI-driven solutions that transform business processes and deliver tangible value.
• PyTorch has launched torchao , a native library that significantly optimises model performance with up to 97% speedup and 73% VRAM reduction for popular AI models like Llama 3, using low-bit quantisation, sparsity, and float8 precision to streamline both training and inference workflows while maintaining minimal accuracy loss.