The AI Chatbot Wars: Google's Gemini vs OpenAI's ChatGPT

This Week

The battle of the bots is on with Google and OpenAI releasing new (and impressive) features last week. You’ll find the highlights and a GPT-4o cheat sheet below. I also share a mini-master class on prompt writing for better accuracy, Section’s 5-day AI crash course and an AI funding update from CB Insights.

Battle of The Bots

At this point, you’re likely a regular user of at least one chatbot, if not a few. There’s no shortage of them: Google’s Gemini, OpenAI’s ChatGPT, Anthropic’s Claude, Microsoft’s Copilot and on. But two of the world’s frontier AI assistants are battling it out for the top spot: Google and OpenAI. 

Google currently has 330 million monthly users (forecasted to reach 1 billion by the end of the year). OpenAI’s ChatGPT is a distant second with 180 million monthly users. Despite ChatGPT’s dominance in the media (and everywhere else), Google has a headstart with its 3 billion Workspace users. It’s only natural these users will gravitate to Gemini over ChatGPT. 

Google is also aggressively investing in its model development and AI-driven solutions (it announced 100 new AI features at its I/O developer conference last week). OpenAI isn’t sitting on its laurels, either. Every time Google threatens to leapfrog ChatGPT’s functionality, they sneak in with impressive developments of their own (their GPT-4o model is the latest example). 

It’s anyone’s guess who will win the top spot, but one thing is clear: this rivalry means we’ll see better and faster AI functionality from both companies. Hold onto your hats! 

Read on for the latest releases from Google and OpenAI.

Google’s I/O Updates in 17 Minutes

Google announced more than 100 new AI features at their 2024 I/O conference last week. Top of the list? Generative media, new models, improved performance, AI bots, multimodal capabilities (that rival OpenAI’s recent release), AI Overview for search, and updates to Workspace. 

You’ll find all 100 feature updates on the Google Blog, or you can read the summary article and watch the video at The Verge.

GPT-4o: Breakthrough AI or Marketing Hype?

OpenAI recently launched GPT-4o, its latest generative AI model. But how do you know what’s truly breakthrough versus marketing hype?

The features that generated the most buzz include real-time voice interaction, the ability to translate speech between 50 languages, and the chatbot’s capacity to perceive emotional cues from users. The real-time voice interaction is particularly impressive, but it’s extremely chatty, so you’ll need patience to stick with it. 

Section’s Chase Ballard sorted through the hype to unearth the breakthrough advancements. His favorites are the multi-modal capabilities (voice, text, vision), faster responses, and real-time voice conversations. Oh, and the fact that OpenAI is giving it away for free.

AI expert Joshua Chang also gives a nice summary of the OpenAI release on his YouTube channel.

Fun fact: the "o" in GPT-4o stands for "omni," representing the model's ability to handle text, speech, and video inputs. 

Your Guide to ChatGPT-4o

Now that you have GPT-4o at the ready, you can use Edward Frank Morris’ cheatsheet guide. 

5 Tricks for Improving ChatGPT Accuracy

It’s true: sometimes ChatGPT lies to you. But Briana Brownell from Descript, has created a mini-master class on prompt writing that improves the accuracy of responses from tools like ChatGPT. Here are her 5 tactics for getting better answers from your large language model:

1. Few-shot prompting: Provide a few examples for the model to learn from.

2. Chain-of-thought prompting: Have the model go through intermediate reasoning steps before answering.

3. Rephrase and respond prompting: Ask the model to rephrase your question and then respond.

4. Self-consistency prompting: Generate multiple responses and select the most consistent one.

5. Knowledge integration prompting: Direct the model to use specific knowledge sources.

If you want the detail, head over to the Descript blog here. 

Master AI in Section’s 5-Day Crash Course

Want to brush up on your AI essentials? Section is offering a week-long AI course in six lessons led by their AI avatar. The AI Crash Course will equip you with the tools and techniques to craft effective prompts, optimize workflows, and maximize personal productivity. 

You’ll also get 1:1 time with (human) AI experts Ed Ortega and Chase Ballard during lecture times and office hours. Section promises you’ll leave with the confidence to apply AI in your work every day (plus a badge to share on LinkedIn). 

What you'll learn:

1. How AI works: its current capabilities, and how to navigate its risks and limitations. You don’t need to be an AI expert – but you do need to “talk the talk” well enough to convince your team.

2. Tackle your to-do list using A:
Audit your workflows to identify where AI can augment your work (aka, high effort and high brainpower tasks). Walk away with a list of tasks that are now AI’s job – not just yours.

3. Unleash AI’s three “personas” on your work:
AI can act as an assistant, a strategist, and a creator, and it’s all about pairing each persona with the right workflow. You’ll identify ways each persona could apply to your work, then build prompts that get the most out of AI.

4. Master the art of writing high-quality AI prompts: Garbage in, garbage out. Learn to write prompts that get a great output from ChatGPT, Claude, or Copilot the first time around.

Sign up for the AI Crash Course at Section. 

Magic, Marketing & Meccano: A Personal Tale

"You need to be more normal."

That's what a former boss once told me when I was being myself at work.

For years, I tried to fit in by hiding my true personality. I grappled with impostor syndrome and went to great lengths to "fit in." But an unexpected conversation with a mentor changed everything. He made me realize the power of bringing my whole, authentic self to work – quirks and all.

I hope that by sharing my own journey, I can inspire others to embrace what makes them unique and look for workplaces that allow them to thrive as their authentic selves.

I'm lucky to have found work I love and colleagues who appreciate my weirdness. If you're interested in finding out how I got there, continue on to the article below. It's a little different from what I usually share, and doesn’t feature AI, but hopefully, you find it entertaining and helpful.

Read on here.