An AI Founder's Vanity, Posturing, and the Peak of Ignorance | Interview with Mengqi, Founder of invoko.ai

Mengqi is a typical representative of one of the most popular types of founders in recent years: she comes from ByteDance, assembled a glamorous team, leveraged the momentum of Agents to tell a compelling story, raised substantial funding, and then pivoted along with technology trends.

Of course, after the exponential progress in AI Coding, like many other founders, she also found herself in a slump due to the overwhelming strength of the models.

What's rare is that, after extensive reflection, review, and learning, Mengqi pulled herself out of the "valley of despair" and became optimistic again, once again believing that software still has great potential.

In the episode, Mengqi shared very honestly her thoughts and mindset changes over the past year. She reflected on several of her pivots: why did decisions that initially seemed highly coherent mostly turn out to be wrong? Why did she later realize that to B wasn’t a promising direction for entrepreneurship? And how did she regain confidence in software after shifting toward the consumer side?

She also engaged in sharp self-criticism: when starting a company, it's easy to be led by vanity and sexy narratives, overlooking what the business and users truly need; and right after securing funding, founders are often closest to the “peak of ignorance.”

In such unfiltered expressions, Mengqi also delivered many gems and provocative statements. For example, there are many “scam companies” in the industry; many vertical Agents are essentially just scenic views for investors; the phrase “XX is dead, XX lives forever” is the “tomato novel” of the AI circle.

While listening, I felt both real and exhilarated, but also genuinely anxious for her. Her honesty even made me wonder if there's anyone left in this industry that she actually cares about (laugh).

But in a time of widespread anxiety and confusion, perhaps this voice that is both human and hopeful is exactly what we need most.

【Human Museum】

Tour Guide: Qu Kai, Founder of 42 Chapters

48th Treasure: Mengqi, Founder of invoko.ai

【Time Machine】

• 02:54 Should all software founders who raised big money last year bow to Manus?

• 03:45 In February or March, I thought I was completely crazy for doing software

• 05:17 AI coding is already so strong — why are there still so few products that actually come out?

• 08:26 Mistakes in the pivot journey:

• 08:45 ① Doing vertical Agents

• 11:40 ② Over-serving professional users

• 12:28 ③ Choosing the wrong scenario

• 17:08 Two structural dilemmas of vertical Agents

• 17:30 Many AI products are just scenery for investors?

• 18:45 Two types of AI software businesses: large R operations & gym model

• 21:18 Can efficiency tools provide emotional value? Steam gave me a lot of inspiration

• 22:48 To me, to B is not a worthwhile entrepreneurial direction

• 24:44 Sharp critique on founder vanity, posturing, and the peak of ignorance

• 28:36 After seeing the antics of US to B, I’m disillusioned about the US startup ecosystem

• 30:21 The stronger AI coding gets, the more I respect software

• 35:58 A C-side product only needs to do one thing to have value

• 39:07 Share a crude but effective user research method

• 44:00 AI might completely颠覆 the mindset of “hitting nails with a hammer”

• 51:02 Spreading optimism: software still has great potential!

• 55:38 Rejecting “the world’s first xxx,” resisting “xxx is dead, xxx lives forever”

• 57:41 The last product was solving market proof problems, now I want to answer a real short-answer question

• 01:01:37 Soul-searching: Do you regret leaving ByteDance to start a company?

• 01:05:00 Entrepreneurship can't rely too much on the left brain

【The gang that made this happen】

• Producer: Chen Pi

• Editor: Chen Pi

• Intro: Mondo Bongo - Joe Strummer & The Mescaleros

• Outro: Sunday Morning - The Velvet Underground, Nico