Don't create ChatGPT competitors. Sam Altman sets a "forbidden zone" for AI companies: We have a "moat" in this area.

At the recent AI Startup School event hosted by Y Combinator in San Francisco, OpenAI CEO Sam Altman was interviewed by Garry Tan. On stage, he said, "Don't try to build what we're building with ChatGPT."

Altman didn't beat around the bush and clearly stated, "What we're going to do is to turn ChatGPT into the best super assistant. We'll add all the features we think are necessary."

Furthermore, he added a statement with obvious emotion: "We'd be very sad if someone said they were starting a business to create an alternative to ChatGPT."

He said, "The core form of ChatGPT is a product line that OpenAI itself will continue to advance. It's not suitable for entrepreneurs to enter through direct competition."

He believes that OpenAI has a "huge first - mover advantage": "We were the first to launch a good product. Now we have a brand, data, memory, and connections. We'll continue to build moats."

Original interview link:

https://www.youtube.com/watch?v=V979Wd1gmTU

What does Altman want to convey?

Caption: There are also discussions on Reddit.

Is this arrogance or a strategic candor?

After all, in the early stage of every technological wave, when the rules are not yet established, everyone can gallop in the wild. But when a player's scale and speed are far ahead, it becomes the de - facto "infrastructure". Instead of competing with the infrastructure, it's better to utilize it.

So, what exactly is the "core business" that entrepreneurs should avoid? Altman believes it's far more than just a chat box. The blueprint Altman painted is an omnipresent and deeply personalized "super assistant". This assistant has memory and can understand you better and better through continuous interaction with you, just like a friend who knows you well.

It will be deeply integrated with all your data sources, connect your devices, and may even have a new hardware form - that's why OpenAI is collaborating with former Apple chief designer Jony Ive. Its interaction mode will no longer be typing but more natural voice. It can even proactively anticipate your needs, handle affairs for you in the background, and only appear when necessary.

From having memory, to connecting data, then to running actively, and finally having its own physical carrier. This is the path OpenAI has drawn.

While defining the restricted area, Altman also pointed out the opportunities. He proposed a key concept: "Product Overhang". That is, the capabilities of the model have reached an unprecedented level, but there are extremely few products that truly make good use of these capabilities. This huge "gap" between capabilities and applications is fertile ground for startups.

The opportunity lies in using powerful reasoning models like GPT - 4o to reconstruct old workflows. For example, AI will not only answer questions but also be able to receive a task like a junior employee, then conduct research, code, execute, and finally deliver a complete solution. This "Just - in - time software" model will subvert the traditional SaaS industry.

In other words, OpenAI focuses on building that omnipresent and increasingly user - understanding general assistant; while outside this main trunk, there are plenty of branches for entrepreneurs to blossom and bear fruit.

Caption: Comments from netizens on Reddit.

Altman: Starting OpenAI was not easy

In addition, in this interview, Altman deeply reviewed the entire process from the difficult early days of entrepreneurship to the establishment of OpenAI.

He said that in 2015, when GPT - 1 hadn't emerged and deep learning was still stumbling in video games and robotic arms, a decision that seemed like a joke in entrepreneurship was quietly made: a few friends gathered around a whiteboard and discussed whether to establish a research institution for AGI.

At that time, the AI world was dominated by DeepMind, and public discussions on general artificial intelligence were still limited to the academic circle. Sam Altman was still the president of YC at that time. He said , "Starting OpenAI at that time was a 'coin - toss decision' because there were too many reasons not to do it."

They had no product, no revenue, and even no reliable direction. There were just a group of people and a whiteboard, aiming to write "papers that could be read". But in this state of "having no reason to believe", Altman and his partners chose to "do the most difficult and most worthwhile thing".

"We said we were going to do AGI. 99% of people thought we were crazy, and the remaining 1% resonated strongly with us." Altman mentioned in the sharing that this is the secret to attracting top - level talents - doing things that seem like "crazy" things can quickly attract "smart people with nowhere to go".

In his view, the companies that can truly attract talents are often not projects that "everyone is doing", but unique missions that "can't be replicated one - to - one".

Altman emphasized many times that YC hears hundreds of "using AI to do XXX" startup ideas in each session, and only a very small number of them are not copying each other. "Truly great companies never start from a consensus among a large number of people."

And it's not easy to bear the cost of being an "outsider". "When Elon sent us an email saying 'Your success probability is 0%', I really wondered if we had messed up that night when I got home." Altman said.

But he also admitted that there's no magic to help you fight against such doubts, only one word: conviction - belief.

This article is from the WeChat official account "BigDataDigest" (ID: BigDataDigest), author: Digest Jun. It is published by 36Kr with authorization.