The Chatbots and Agents Are Going To Merge
Could the AI labs intuit your next move and accomplish it for you? That seems to be where they're heading.
There are two core generative AI uses today: Chatbots and agents. The bots make sense of information. The agents take control of your computer and get things done for you. The two seem distinct, but they’ll likely soon merge.
When you use a chatbot to research hotels, experiences, or companies, the bot, with your permission, will one day almost certainly take control of your browser to reserve a room for you, email a service provider for more information, or even purchase a company’s stock. The bot will suggest these actions as you chat with it. Then, upon approval, it will take command and get to work.
Today’s AI labs, already seeing the value of autonomous actions in coding, view you leaving the chat window as somewhat unnecessary, and they seem ready to complete the circuit. They want their chatbots, effectively, to intuit your next move and accomplish it for you.
OpenAI, for instance, is building a ‘super app’ for autonomous coding in Codex, but that concept may soon broaden beyond code. The updated Codex app includes an autonomous coding tool, a browser it can make use of, and an instance of ChatGPT. But the ambition extends to the core ChatGPT application itself.
“We are busy bringing ChatGPT to Codex so that we can bring Codex to ChatGPT,” said OpenAI engineer Thibault Sottiaux on X recently. “One day this will make sense.”
When you’re researching a service, let’s say a child’s birthday party entertainer, ChatGPT might, for instance, suggest some possible hires and then compose and send emails to them inquiring about availability and price. If the clown or magician meets your specifications, the bot could then confirm them, and maybe even run a background check. Just set the parameters, approve the language, and move on to your next chat.
Chatbot actions may soon extend far beyond today’s conceivable use cases, potentially to areas like stock trading. Robinhood, for example, just built an integration that allows users to delegate their trading decisions to AI bots. But soon, when researching a company in a chatbot, the bot might offer to buy shares for you. It might even help shape an investment strategy, get a budget, and execute it while you wait.
Boris Cherney, the head of Claude Code at Anthropic, told me on Big Technology Podcast last week that the company is experimenting with more proactive versions of its Claude chatbot. “We’re trying all these different experiments,” he said. “There’s some stuff that we’re trying that’s like this.”
To be sure, this product vision will only be realized if people trust the AI labs to do the right thing with their data. And given the polling right now, that could be tough.
But chatbots that go beyond the chat window are certainly on their way. And should they catch on, they will have a profound impact on the web and the services built around it, potentially becoming central hubs for all computing. Outside websites and apps would be there for fact checking and triple confirming certain details only.
The term ‘super app,’ though initially sounding like a misnomer, is therefore the right way to describe the ambition. In places like China, a super app is a social network, payments tool, and ride hailing service, all in one. The new agent-chatbots could sync with all the same programs, and potentially do much more. This is the bull case for generative AI, and it’s not a totally inconceivable story. It’s likely the narrative OpenAI and Anthropic will take to Wall Street en route to trillion-dollar IPOs.
Investors are watching this fast growing tech company (sponsor)
🚨 No, it’s not the publicly traded tech giant you might expect… Meet $MODE, the disruptor turning phones into income generators.
📲 Mode’s 32,481% revenue growth ranked them #1 on Deloitte’s list of fastest-growing companies in software. They aim to pioneer “Privatized Universal Basic Income“ powered by technology, not government, and their EarnPhone has already helped consumers earn & save $1B+. However, their current share price is set to change on 5/29.
⏳ Time is running out: $0.50/share pricing ends at 5 PM ET today.
What Else I’m Reading, Etc.
Runaway token costs are starting to worry big corporations [WSJ]
Now Dario and Sam say AI might not take our jobs after all [Inc]
So, what is AI actually doing to jobs? [Kyla Scanlon]
Memory stocks are ripping amid the AI boom [WSJ]
Anthropic passes OpenAI to become the world’s most valuable AI startup [New York Times]
Join the Big Technology AI Summit in San Francisco on June 18
Dylan Patel will be there too. Get your ticket at summit.bigtechnology.com
This Week On Big Technology Podcast: Predicting the SpaceX, OpenAI, and Anthropic IPOs — With Dick Costolo
Dick Costolo is the ex-CEO of Twitter and managing partner at 01 Advisors. Costolo joins Big Technology Podcast to look ahead to the SpaceX, OpenAI, and Anthropic IPOs, looking at whether the offerings will be successful, who should go first, how the companies differ, and where the capital will come from. Tune in for the second half where we discuss the state of Meta, Costolo’s Twitter memories, and Silicon Valley’s ‘Permanent Underclass.’ Tune in for a fun, in-depth discussion of the most important pending business event in decades.
You can listen on Apple Podcasts, Spotify, or your podcast app of choice
Thanks again for reading. Please share Big Technology if you like it!
And hit that Like Button, or have a chatbot do it for you…
My book Always Day One digs into the tech giants’ inner workings, focusing on automation and culture. I’d be thrilled if you’d give it a read. You can find it here.
See you next week!
Disclaimers
Please read the offering circular and related risks at invest.modemobile.com. This is a paid advertisement for Mode Mobile’s Regulation A+ Offering.
Mode Mobile recently received their ticker reservation with Nasdaq ($MODE), indicating an intent to IPO in the next 24 months. An intent to IPO is no guarantee that an actual IPO will occur.
The Deloitte rankings are based on submitted applications and public company database research, with winners selected based on their fiscal-year revenue growth percentage over a three-year period.
Sounds like Jensen Huang unveiled exactly this in Taiwan tonight!
The merge is real but the constraint isn't capability, it's the approval surface. Every example here routes through a human gate: approve the email, approve the trade, confirm the magician. The product battle is whether that gate stays legible when the agent moves faster than a person can actually verify, because a rubber-stamped approval is just an agent with no brakes.