Mache keine Konkurrenzprodukte zu ChatGPT. Sam Altman markiert für KI - Unternehmen "Verbotszonen": Hier haben wir eine "Schutzmauer".

At the recently held AI Startup School event by Y Combinator in San Francisco, OpenAI CEO Sam Altman was interviewed by Garry Tan. On stage, he said: "Don't do what we're planning to do with ChatGPT."

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

Moreover, he added a sentence with strong emotion: "We'd be very sad if someone said they were starting a company to create a replacement for ChatGPT."

He explained: "The core form of ChatGPT is a product line that OpenAI itself will further develop. It's not suitable for entrepreneurs to directly compete with."

He said OpenAI has "a huge head - start": "We were the first to bring a good product to the market. Now we have brand recognition, data, memory, and connections. We'll continue to work on our moat."

Original interview text:

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

What does Altman mean?

Caption: There are also discussions on Reddit.

One doesn't know if this is arrogance or strategic openness?

After all, in the early stage of every technological wave, when the rules aren't set yet, everyone can roam freely. But when a player is far ahead in size and speed, it becomes the de - facto "infrastructure". It's better to use this infrastructure than to compete with it.

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

It will deeply integrate all your data sources, connect your devices, and even have a new form of hardware - that's also the reason for the cooperation between OpenAI and the former chief designer of Apple, Jony Ive. The way of interaction won't be typing anymore, but natural language. It can even predict your needs and complete tasks for you in the background and only appear when necessary.

From memory, data connection, to active operation, and finally to its own physical form. That's the line OpenAI has drawn.

While setting the no - go zones, Altman also pointed out the opportunities. He brought up a key concept: "Product Overhang". That means the capabilities of the model have reached an unprecedented level, but there are very few products that really use these capabilities. This large "gap" between capability and application is fertile ground for start - ups.

The opportunity lies in using powerful inference models like GPT - 4o to reinvent old workflows. For example, AI can not only answer questions but also take on a task like a young employee, then conduct its own research, programming, execution, and finally deliver a complete concept. This "just - in - time software" approach will turn the traditional SaaS industry upside down.

In other words, OpenAI focuses on the omnipresent, ever - better - understanding universal assistant. But outside this main line, there are still many branches where entrepreneurs can do successful business.

Caption: User comments on Reddit.

Altman: Founding OpenAI wasn't an easy task

Moreover, in this interview, Altman traced the entire journey from the difficult beginnings of starting a business to the founding of OpenAI.

He said that in 2015, when GPT - 1 didn't exist yet and deep - learning research was still in its infancy and focused on video games and robot arms, a decision was made that sounded like a founding joke: Some friends sat around a whiteboard and discussed whether they should found a research institute for Artificial General Intelligence (AGI).

At that time, the AI world was dominated by DeepMind, and the public discussion about AGI was limited to the academic world. Sam Altman was still the president of YC at that time. He said , "Founding OpenAI was a 'coin - flip decision' back then, because there were so many reasons not to do it."

They had no product, no revenue, and even no reliable direction. There was just a group of people and a whiteboard, and they wanted to "write papers that someone could read". But despite this "groundless approval", Altman and his partners chose "the most difficult and valuable thing".

"We said we wanted to create AGI. 99% of people thought we were crazy, and the remaining 1% felt strongly attracted." Altman said in his speech that this was the secret to recruiting top - notch talent - to work on things that seem like they're from "crazy people" to quickly attract "smart people who don't fit in anywhere".

In his opinion, it's not the projects that "everyone is working on", but the unique missions that you "can't copy 1:1" that can really attract talent.

Altman emphasized several times that in each round of YC, he heard hundreds of business ideas with the buzzword "AI for XXX", and only very few of them were not just copies of each other. "Real big companies never start with an opinion that many people share."

And it wasn't easy to bear the cost of "being different". "When Elon wrote us an email and said, 'Your chance of success is 0%', I really wondered if we had messed it up." Altman said.

But he also admitted that there's no magic to fight against such doubts, just one word: Conviction.

This article is from the WeChat account "Big Data Digest" (ID: BigDataDigest), author: Wenzhaijun. 36Kr has obtained permission for publication.