[ad_1]
Right now, there is hardly any programmer in the world who does not use an AI-based co-pilot in some way. But using GitHub Copilot or Cursor.AI to ask technical questions and get help debugging may be just the beginning. AI coding may one day include agents that can write programs themselves based on a natural language prompt. Such programs may replace human engineers.
AI coding startups that can generate code from a natural language prompt include Replit and Bubble, among many others.
Eventually, some venture capitalists believe companies will hire fewer human engineers and have every human running AI crypto agents. “It’s not pie in the sky. It’s in the near future, but it’s not today,” VC Corinne Riley, a partner at Greylock, said on stage at TechCrunch Disrupt last week.
She added that coding assistants are already allowed during the technical coding interview with potential employees at several Greylock portfolio companies.
However, she doesn’t think AI agents should be used to replace human engineers in startups to truly save money. In the seed stage, “what you’re doing is you’re building the foundations of the company, right? So, if you’re making big engineering trade-offs at that stage, it’s probably not the right decision. These are decisions you can make in the future,” she said.
However, cash management is also why young startup engineers should be used to help program AI as much as they can right now, VC Elizabeth Yin, co-founder and general partner of Hustle Fund, said on stage.
“One of the main challenges in the early stages is that you don’t know exactly what problem you’re solving and exactly what your ICP (ideal customer profile) is and exactly what they need. So you’ll end up eliminating a lot of work,” Yin said. The faster you can go, and the faster you can iterate, the better for learning quickly.”
She believes that early-stage startups should be open to any tool that allows founders to quickly put together product models to move faster, even if they have to be rebuilt more carefully later. “I would actually be all for it if it meant you could learn it much faster,” she said.
This is in contrast to the days before artificial intelligence, when every pilot had to be coded by someone with the skills. Today, an engineer can ask for a model, use some AI corrections, and take a peek.
The agreement was made with early-stage VC Renata Quintini, co-founder of Renegade Partners.
“If it’s finding out or testing product-market fit, you have to use that leverage, but I don’t worry about optimizing that at the seed stage,” she said on stage.
Interestingly, with startups founded in 2024 using AI development processes, we could be witnessing the seeds of the first future AI workforce. The first to get AI agents as co-workers are the programmers themselves. It’s a thought that’s ironic and prophetic in equal measure.
[ad_2]