The author of this article shares his experience and experience of failing AI entrepreneurship, from the challenges of technology entrepreneurship, the prerequisites for entrepreneurial success, to the half-hearted entrepreneurial dilemma, to the potential of the AI+2C market and platform dividends, providing valuable lessons and enlightenment for readers who are interested in AI entrepreneurship.
AI entrepreneurship fails, you can share lessons learned privately…
Entrepreneurial failure experience series of articles: Brother, find a class! Don’t be superstitious about AI entrepreneurship, I regret leaving my job Let’s talk about how I was “deceived” by AI entrepreneurship and how I failed
After May, I really couldn’t see the possibility of success in my AI entrepreneurship, so I decided to give up, and the reasons for the failure were also summarized and sorted out for everyone, so as not to fall into the pit: technology entrepreneurship = stupid
First of all, unless you are a top 10 in the industry, technology entrepreneurship, including AI entrepreneurship, is a very ridiculous thing, don’t think that you have the latest technology in the industry, you should come out to start a business, look at the opportunity, don’t be stupid, there is no good fruit outside!
PS: Offensive, I know two industry giants, in fact, I didn’t do what I thought was entrepreneurship, they were selling courses… Two premises, one possibility
Second, there are two premises and one possibility for the successful entrepreneurs around me, in order, everything is ready, desperate, and lucky.
Those who are ready for everything, generally speaking, are eating the dividends of the previous company, most of them have already run the model in the previous company, or take various resources of the former company to start a business, this path is very suitable for the person in charge of the business side;
Desperate people are generally obsessed with something, must do it and can no longer find other upward channels around them that can satisfy their ambitions, they are easy to go bankrupt to start a business, and there are also successful after several failures, this group of people are very ruthless to themselves…
Lucky people, I belong to my type, stupidly taking a product or technology to start a business, the product itself is not selling well, but in the process of doing an outsourcing project, I suddenly found inspiration to become bigger and stronger!
Therefore, there are actually only two prerequisites for survival: either there is a lot of accumulation before starting a business, or funds or orders; Or you can only be lucky to do a project, and the project is successful… Three hearts and two minds, opportunities everywhere
Third, out of the company system, you will find that there are opportunities to make money everywhere, take me as an example: first, I am doing self-media work, publishing articles, making small videos, doing live broadcasts, and even the theme is dispersed into two major topics: management and AI; Secondly, I have my own AI+ managed products, and I have invested a lot in this area, and I have always been blood transfusion creation, but the final result is that it does not sell well; Then, in order to survive, we had to take on a lot of AI projects, this job was very pre-sales, he ate a lot of my energy, and I was often prostituted for free in the process; Finally, various AI opportunities were found in the middle, and many times I wanted to intervene, but I immediately found that there was a certain threshold and it was difficult to go deeper;
In the process of communicating with a certain boss on the weekend, I was a little discouraged, doing self-criticism, which means I wanted too much, but in the end I didn’t do it well…
The big guy, first of all, affirmed that there is indeed a problem with half-heartedness, but he is also expressing: many times, entrepreneurs have no way to be essential, they can really only want both and want, if they can’t do it, it will be difficult to survive.
Therefore, he believes that half-heartedness is not too much of a problem, the most important problem is that there is indeed no line explosion in the process, and the core is still a bad environment and bad luck.
In short, although the AI era is coming, the current threshold for entrepreneurship is actually very high.
And for me, I have to think about the follow-up plan…
AI + 2C dividends are huge
Recently, two voices have appeared around me:
- Find a class;
- To do 2C, it can be AI or management, preferably AI;
For example, my good friend Shen Yang suggested that I go to work before, and recently he also suggested that I do AI + 2C. Because the logic of AI 2C is completely different from that of 2B, from a certain level, 2C is easy to make money and has a low threshold!
First of all, with the explosion of DeepSeek, the cost of AI education in China is extremely low, and most people are anxious and even willing to take the initiative to seek payment!
From the perspective of industry layout, it is also a huge dividend:
- ByteDance’s capital expenditure in 2024 will be 80 billion yuan, doubling to 160 billion yuan in 2025 (including 90 billion computing power and 70 billion IDC), and the monthly active users of 16 types of AIGC models of “Doubao” will be nearly 60 million;
- Alibaba has invested more than 380 billion yuan in three years to build AI+ cloud, and “Qianwen 3” and 200+ Tongyi open source models (more than 300 million downloads) support retail, manufacturing, media and other scenarios, and revenue has grown by triple digits for seven consecutive quarters.
- NetEase “Reverse Water Cold” used AIGC to reduce character production costs by 90%+, game DAU jumped 47%, Q1 net profit increased by 34.9% to 10.3 billion yuan, and the stock price approached an all-time high;
- Tencent’s Q1 capital expenditure increased by 91% year-on-year to 27.5 billion yuan (90% was invested in AI), and the delay in self-developed “mixed yuan” advertising shrank to one-tenth of the original one; The AI assistant “Yuanbao” borrowed the WeChat ecosystem to break 20 million DAU in two months;
Because of the boom of the industry, the development of various AI applications has been promoted, especially in consumer-level scenarios such as advertising, game levels, short video editing, copywriting and customer service:
AI’s natural advantages in unstructured data processing and personalized generation are redefining content creation with more than 10x efficiency.
For 2B: The “complexity moat” on which traditional enterprise software depends is being quickly filled by AI’s “general intelligence + modular calling”.
In other words, in the past, if a company did something a little, it would do digital transformation, which involved a large budget, but AI seems to be changing all of this (and may be making it up), and it will give companies an illusion:
If an AI can help me handle reimbursements, do financial reconciliation, and output audit reports, then why should I buy a lengthy and complex ERP system? I only pay for the results.
AI is omnipotent, and the expected hustle and noise of traditional ERP and CRM systems is rising…
This is actually adapted to Sequoia’s understanding of the development of AI at the closed-door meeting: the next round of AI will not sell tools, but benefits.
To sum it up: AI has made the appetite of enterprises more satisfying, which is good for 2C and bad for 2B. The result is as described by his friend Shen Yang:
- Small and medium-sized enterprises: AI turns the once “unaffordable, unusable” capabilities into a plug-and-play digital workforce, and solo entrepreneurship can also outsource finance, legal affairs, customer service, and operations.
- Large enterprises: AI reshapes the flow of information, decision-making and execution, helping to shift from a manpower-intensive process-driven to a lightweight, automated operation system.
- Software vendors: Whether to sell tools by license or transform into AI-native and results-based intelligent services has become a life-and-death decision.
On this basis, various major manufacturers quickly laid out the Agent entrance and generated huge dividends, such as the typical application of byte buttons and Tencent’s IMA knowledge base.
How can product managers do a good job in B-end digitalization?
All walks of life have taken advantage of the ride-hail of digital transformation and achieved the rapid development of the industry. Since B-end products are products that provide services for enterprises, how should enterprises ride the digital ride?
View details >
Platform dividends
We have actually talked about the path of ordinary people using AI:
- Use a tool like DeepSeek directly;
- Use the buckle agent platform to build an AI assistant;
- Formation of Agent platform;
In fact, the implementation cost of the above is not high, and large platforms such as Byte and Tencent quickly provide various platform-level services, and smart and sensitive programmers will build various agents based on these platforms such as buttons.
In fact, this pattern is very clear, such as dragging and dropping buttons to generate general agents, such as Cursor to generate basic code, and ComfyUI to generate the design drawings that designers want.
Their essence is model + workflow, further evolution is a general platform like Manus, and taking a step back is the transformation of traditional digital cost reduction and efficiency increase.
It’s just that applications like Button and ComfyUI can be considered as a kind of low-code platform, an assistant for professional people, and it is still very difficult for ordinary people to use, although it is difficult but it can’t stand everyone’s curiosity, so a huge business opportunity has been derived:
Teach ordinary people to use button-type platforms
For example, my friend Lao Bao Agent, he has built a large number of applications based on buttons:
These agents based on buttons are hard to say, but they must be very decompressing!
In terms of income, he has surpassed the average small company, and the profit is very considerable!
Many friends actually have a question: can this kind of thing with such a low threshold make money?
Here is the right guide’s thinking: the gold content is not good, the problem is the most important, the evaluation should be handed over to the market, and the market is willing to pay, that is right.
So, the real point here is: don’t despise or be jealous, as for me, the question here is: do you want to do AI + 2C?
epilogue
In fact, I thought about this a few years ago, when I wrote a management course and wanted to sell it:
But in the end, I couldn’t type 98.8 advertisements that let you grasp the essence of management, and finally sent them out for free…
The core contradiction here is: I think 2B and 2C are in conflict!
No matter what 2C does, it will make people feel like they are cutting leeks, and I cherish feathers a little, after all, the official account has not even played an advertisement for so many years.
There is a trade-off between word of mouth and brand, if you sell classes and cut leeks, it may not be easy to return to doing difficult things.
If you just want to make some money to eat, it may be a good choice to do self-media selling courses.
But for the time being, I’m still going to find a class, and I’ll take a look in two years if it’s 2C…
