When contemplating where OpenAI is heading in the next five years, Sam Altman’s history and demeanor are telling indicators. As a prominent figure from Y Combinator, few CEOs have had the depth of exposure to startups and businesses that he had during his tenure there. Listening to most public interviews with Altman from Lex Friedman to Oprah, suggests that Sam envisions a much grander scope beyond just ChatGPT and the quest for a general artificial intelligence. He appears to have ambitions reminiscent of Google’s scale; he aims to be a major player in the tech industry on a global scale.
Take, for example, the integration of ChatGPT into various applications and platforms such as a search engine. Many doubted that OpenAI would venture into areas dominated by big entrenched monopoly tech companies, yet they’ve made significant strides in expanding their AI capabilities. The continuous improvements in their models have made them competitive, if not superior, in certain aspects compared to industry giants like Google.
So, what can OpenAI do to continue drawing attention away from competitors like Google? Given OpenAI’s substantial resources with apparently both money and “compute to burn”, there’s no reason they can’t launch larger, public-facing offerings.
One aspect that is unique to OpenAI, is that they have hundreds of devs soaking in AI/LLM constantly. They know how to use and apply the power of AI to extract a desired result. They know exactly how to use ChatGPT’s coding prowese to generate new services and features almost over night. It took them less than a year to go from ChatGPT 3.5 to SearchGPT. Some of the ideas below are now so trivial, that it would be dumb if OpenAI didn’t do them.
Sam Altman interview with Lex Fridman – March 2024
Opening up my crystal ball (that has been polished after 45yrs in tech)
New offerings need to focus on increasing time on site and integrating OpenAI into people’s daily routines. The longer OpenAI can keep ChatGPT, SearchGPT and their logo on peoples screens, the more more successful they will be.
- A free email service: yes, it is tired trope, but so was Gmail when they cloned Hotmail and Yahoo mail. Yet it became a dominate player not because of quality-of-service, unique offering or positioning – it only worked because of well Google. It was simply the power of Google search that shoved Gmail into the psyche of users. OpenAI can leverage ChatGPT offerings to bake AI into an email service and offers things Gmail/Outlook don’t (ease of use being primary).
It’s boring, it’s old, but nothing takes eyeballs like email and messaging services.
OpenAI could position this as totally counter-opposing to Gmail:- Privacy First: “We don’t index, read, or scan your email for targeting adverts“. I’d hit privacy hard where Google is vulnerable on privacy.
- Attack Phishing issues excessively. Top complaint against Gmail is phishing emails. Given that most connect to Gmail in a browser, clicking on phishing emails is a significant risk. OpenAI can promote privacy and security.
- Attachments. Gmail limits you to 25mb. Sheeze, I have single photos that are larger than that these days.
- Simple to use Rules system. Gmail is a mess of folder and categorization nonsense. You miss important stuff because it ends up in some bizarre folder without asking.
Lastly, bake in a direct “messaging” service – because, why not? Dead easy, convenient, and creates user lockin. Facebook wouldn’t be where it is without messenger and WhatsApp.
- A web browser: While acquiring a company like Mozilla would be ideal (best for end users), it’s more likely we’ll see a Chromium-based offering. When you look at what some of the other browsers like Opera are doing with AI, it is intriguing. OpenAI could create something remarkable with integrated ChatGPT capabilities.
- Enhancing the search experience: Developing a standalone domain and rebranding their search tools, possibly adding services like maps. I did a stand alone post on this elsewhere. The SearchGPT offering is stunning.
- Advancing autonomous agents: IMHO: Since previous attempts like the GPT Store didn’t gain substantial traction or land with the public quite the way we thought they would – most expect OpenAI to rebrand and push the agent paradigm forward. Autonomous agents are rapidly evolving; hopefully, OpenAI can perfect them.
- Entering the web advertising market: Ultimately, subscription models have limitations, but advertising offers limitless revenue potential. Capturing the “webmasters” accross the web with a new model, would do more for the future of OpenAI than anything else they could do in the next five years. I honestly have not heard anything out of Altman or OpenAI to think he gets this at a required visceral level.
- Portal Bits and Pieces: calendar integration, todo…etc
- Docs/Spreadsheets/etc: Are hard and offer little reward other than eyeballs. I don’t think they’ll do it, but gosh, leveraging AI here is fascinating opportunity. Not sure they will do it.
- Expanding into cloud services: OpenAI might offer AI-optimized cloud computing services, competing with the likes of AWS and Azure, leveraging their expertise in AI workloads. Meh whatever.
- Investing in hardware: Developing specialized AI chips or hardware to optimize their models’ performance and reduce dependency on external suppliers. There has been noise here, but I don’t think hardward ambitions should be more than water cooler talk.
Potential Challenges:
- Relations with Microsoft: In the next 12 to 18 months, tensions may rise as OpenAI launches more products that directly compete with Microsoft’s offerings.
- Regulatory hurdles: As AI technologies become more pervasive, OpenAI may face increased scrutiny from regulators concerned about privacy, security, and ethical implications. Fortunately for progress, Govt’s are too slow, bloated, and clueless about tech to do anything significantly damaging to the techstack here.
