Thinking

Last Month in AI:
Strawberries

Published
October 8, 2024

September served up a juicy surprise in the world of AI: OpenAI’s new ‘o1‘ model, codenamed ‘Strawberry’ was released. OpenAI o1 isn’t just another AI model; it’s pitched as a crucial second step towards true ‘AGI’ – silicon-based general intelligence on par with humans. (FYI, there are apparently just 5 steps to AGI, and OpenAI’s CEO Sam Altman says internal progress to step 3 is encouraging, and AGI, and then ASI (artificial superintelligence) is now just a few thousand days away. Tick tock, humans). 

By: Paul Marsden, Digital Strategist SYZYGY GROUP

Cognitively, what’s special about OpenAI o1 is that it mimics our brain’s ‘System 2’ reflective reasoning ability, adding a new mode of thinking to ChatGPT’s ‘System 1’ stimulus-response mode. This matters because System 2 thinking is the hallmark of human intelligence – the ability to reason, and work out what to do, when we don’t know what to do. 

In tests, o1 appears rather good at working out what to do when it doesn’t know what to do, acing novel math problems and exhibiting PhD-level reasoning to solve new conundrums. To do this, it employs ‘chain-of-thought’ processing, working out how to answer before answering. In contrast, GPT-series models simply jump straight to the answer based on similarities to past problems. Both kinds of thinking, System 1 thinking and System 2 thinking are important for human intelligence. Now, with o1, AI mirrors the mental architecture of the human mind, and can do both. TL;DR: OpenAI has created a silicon simulation of the human mind. Feeling obsolete yet? 

True to form, OpenAI’s o1 announcement came with corporate drama. They’re switching to for-profit status, raising funding to value the company at £150bn, giving CEO Sam Altman a 7% stake worth $10bn+. Meanwhile, CTO Mira Murati and Chief Research Officer Bob McGrew announced their departures. Are they off to plant AI strawberry seeds for their own startups, or just freeing up time to enjoy humanity’s final moments?

Three Big Implications for Marketing

What does o1-style intelligent technology mean for marketing? Nobody knows, but the new phase of AI has the potential to be highly disruptive in an already disrupted industry. 

  1. Rise of the Agents: Forget AI co-pilots and assistants; OpenAI o1 makes true AI ‘agents’ possible. These agents can act autonomously, solving problems, making decisions, and getting things done. Salesforce and Hubspot have both since announced agents as the future for AI-optimised businesses. Expect marketing agencies to be increasingly populated by teams of AI agents collaborating on your brief. And expect them to talk about AI agents ad nauseum. You have been warned. Microsoft’s agentic upgrade to Copilot hints at a near-term future where AI can do your job, not just assist with it. News leaked that Microsoft was training its AIs on LinkedIn data, and that the Redmond giant is investing in nuclear power to fuel armies of AI worker agents. 
iframe

Mit dem Laden des Videos akzeptieren Sie die Datenschutzerklärung von YouTube.
Mehr erfahren

Video laden

PHNjcmlwdCB0eXBlPSJ0ZXh0L2phdmFzY3JpcHQiPndpbmRvdy5TeXpDb25zZW50UHJveHlbInZpZGVvLWRmMjc4Yzc1LTI1YzAtNDExOC1iYjEwLTE1MDdjZjc0MGQ3YiJdLmNvbnNlbnQoKTs8L3NjcmlwdD4=

2. True Marketing Automation:

With reasoning capabilities, o1 enables genuine marketing automation. A harbinger of this future is Hollywood studio Lionsgate’s deal with AI creative automation company Runway, potentially making storyboard and VFX artists obsolete. James Cameron, who directed Terminator, yes Terminator, joined AI lab Stability’s board – if Skynet becomes self-aware, we’ll know who to blame. Meanwhile, Amazon and Google launched new marketing automation tools, and Klarna’s CEO announced plans to replace both humans and expensive SaaS solutions with in-house AI-developed alternatives. 

3. Creative Destruction:

Smarter o1 technology will make new things possible in marketing, replacing older ways of doing things. In September, Mondelez announced a new smart AI platform that will personalize ads so effectively, you’ll crave an Oreo before realizing you’re hungry. Google released a smarter version of NotebookLM that transforms dull whitepapers and brochureware into (somewhat less dull) podcasts. And Qatar Airways released a new ad campaign that uses AI deepfake tech to let you star in a glossy mini-movie (just ignore the fine print about perpetual rights to your digital doppelgänger). It’s a brave new AI marketing world. 

iframe

Mit dem Laden des Videos akzeptieren Sie die Datenschutzerklärung von YouTube.
Mehr erfahren

Video laden

PHNjcmlwdCB0eXBlPSJ0ZXh0L2phdmFzY3JpcHQiPndpbmRvdy5TeXpDb25zZW50UHJveHlbInZpZGVvLWViNzQ0ZTJlLWNjNWYtNDQwZi05ODVhLTdiOTU3ODQ4OTVkYiJdLmNvbnNlbnQoKTs8L3NjcmlwdD4=

Strawberry Sunset

As a psychologist interested in the impact of AI on humans, I see OpenAI o1 as a landmark release, potentially representing a greater leap towards true machine intelligence than the earlier GPT-series models. While GPT models respond, OpenAI o1 reflects before responding—a crucial distinction. This is exciting, and a little anxiogenic to be honest. My p(doom) number—the likelihood of silicon-based AI replacing us—has historically been low, but with o1 it has now spiked. On the other hand, what a privilege it is to be right here, right now as humanity steps into a new Intelligence Age.

As we mentally prepare to pass on the baton of intelligent life to a new silicon superintelligence, there seems only one fitting response: Ask o1, the planet’s smartest AI, how to craft the perfect strawberry daiquiri, then savour the strawberry sunset of humankind. Cheers to obsolescence!

Let’s talk about AI
Dr. Paul Marsden
Dr Paul Marsden Digital Strategist
SYZYGY GROUP
On this page