- Building AI Agents
- Posts
- GPT-5 is almost here: why it could change the world
GPT-5 is almost here: why it could change the world
Plus: a self-funded agent startup sells for $80 million, 10 strategies OpenAI uses to create agents, and more
Michael Cunningham
June 23, 2025
Welcome back to Building AI Agents, your biweekly guide to everything new in the field of agentic AI!
In an internal memo today, Amazon CEO Andy Jassy revealed that Amazon is massively scaling its adoption of generative AI. $AMZN
Corporate Layoffs Coming: Jassy warned that as AI and agent based tools roll out,
“we will need fewer people doing some of the jobs being done today,
— Sonny Day (@SonnyDayCreates)
12:15 AM • Jun 19, 2025
So does this mean we can get Prime for $12.99 a month again?
In today’s issue…
Why GPT-5 could be transformational
A self-funded agent startup sells for $80 million
10 strategies OpenAI uses to create agents
The lethal trifecta for agents
…and more
🔍 SPOTLIGHT
OpenAI CEO Sam Altman | Source: YouTube
GPT-5 is just around the corner, and—if it’s anything like the last two—it could change AI forever.
In an episode of the OpenAI podcast released last week, CEO Sam Altman revealed that OpenAI’s highly anticipated GPT-5 would be released sometime this summer. According to some early testers, the model is a substantial advance on the company’s previous large language models (LLMs).
But will it be truly revolutionary, or just evolutionary? History gives us some clues.
OpenAI’s release of GPT-3 in May 2020 was widely seen in the tech industry, even at the time, as an eye-opening moment about the power of LLMs. Ten times larger than any model that had come before, it shattered records on a wide range of benchmarks, wrote text often indistinguishable from that written by humans, and showed the first hints of real reasoning powers. Although the launch of ChatGPT—powered by a souped-up version of GPT-3 called GPT-3.5—in November 2022 made the general public aware of the LLM revolution, many insiders saw it coming when they first read the GPT-3 paper two and a half years earlier.
GPT-4’s arrival in March 2023 provoked just as much excitement, but more importantly, it had an immediate impact: within just weeks of its release, the first AI agents emerged. Although earlier models like 3.5 had been capable of powering simple conversational apps like ChatGPT, GPT-4 was the first to achieve the kind of reasoning and long-term coherence necessary to power intelligent apps that could accomplish real-world tasks. Clever hackers put together AutoGPT, BabyAGI, MetaGPT, and more—and the AI agent field as we know it today was born.
GPT-3 set off the AI boom, and GPT-4 set off the agent boom. Will GPT-5 have the same level of impact?
It’s hard to say. OpenAI was the unchallenged king of LLM makers when GPT-4 was released; now, it is locked in close competition with Anthropic, Google, and Meta, and many believe that the low-hanging fruit of LLM advances has already been picked.
Nevertheless, OpenAI has a history of finding new rabbits to pull out of its hat—and the world hasn’t been the same since. I wouldn’t count on GPT-5 being just another gradual step forward.
🎉 JOIN THE COMMUNITY & COURSE
It’s official!
Why join the community? 🤔
Learn to build agents in our n8n masterclass, no coding needed
Copy our step by step agent workflows
Weekly strategy/tech support calls + private builder community
Exclusive, interactive talks with agent founders & builders
Join before the price increases!