😺 Claude 3.5 Sonnet vs. ChatGPT-4o

PLUS: Ultimate Claude vs. ChatGPT faceoff

Welcome, humans.

There’s a new social media app called Butterflies that has users create AI characters who then post photos and interact with other accounts. No offense, but we can’t think of any worse way to spend our time.

If the AI bubble pops, I’m blaming Butterflies and the investors who gave them $4.8M to build the only social media app faker than Instagram (Butterflies literally brands itself as not real life). We woke up and chose violence today!

Here’s what you need to know about AI today:

Anthropic released Claude Sonnet 3.5, and it’s on par with GPT-4o.
Claude Opus 3.5 is coming later this year.
Forbes sent a letter to Perplexity accusing it of copyright infringement.
YouTube will require creators to disclose when videos include AI.

Claude Sonnet 3.5 is the new #1 chatbot in the world.

Anthropic released its best AI model, Claude Sonnet 3.5, and according to most reports and benchmarks, it’s better than any other chatbot, including ChatGPT-4o.

The new model is 2x faster than Claude 3 Opus (Anthropic’s previous best), noticeably better at coding and writing, and significantly stronger in visual capabilities (like explaining a chart).

Here are some amazing demos:

Making a Mario clone using just geometric shapes (link).
A complicated physics question that stumped GPT4o (link).
Coding a fully functional Mancala web app in 25 seconds (link).
Creating an 8-bit style crab using SVG (link).
Transcribing genome sequencing milestones into JSON (link).
Generating diagrams of the chip fabrication process (link).

You can try it for free here.

Since benchmarks don't always reflect usefulness in daily tasks, we compared Claude 3.5 Sonnet with ChatGPT-4o in a few tests.

First impression: Claude Sonnet 3.5 is noticeably faster than ChatGPT-4o. It also seems to just get what we’re looking for better.

Claude Sonnet 3.5 is the stronger writer. From our tests, it’s more human-like, concise, and creative (don’t worry, it’s not writing this newsletter, we pinky promise)!

Many software engineers report it can solve coding issues that GPT-4o can’t.

It’s also better at handling large documents. The 200k+ token context window means you can search and summarize up to 350 pages of text.

Even though Claude Sonnet 3.5 is technically the smarter model, it lacks some useful features that ChatGPT-4o has, like an image generator, Voice Mode, GPTs, Memory/Custom Instructions, and conversation sharing.

Coming soon:

Artifacts: Whatever you’re working on with Claude—code, text documents, designs—will appear on the right side of your screen in a dedicated window.
- This upgrade will make it feel less like a chat and more like you're tag-teaming a project in real time.
Memory, aka Claude can remember things about you from chat to chat.
Claude 3.5 Opus, a model that will be even better than 3.5 Sonnet.

Here’s our take: Anthropic has shown it’s neck-and-neck with OpenAI in terms of model quality. Claude 3.5 Sonnet is seriously impressive, and we’ll be using it alongside ChatGPT-4o in the coming weeks.

That said, most of the general public hasn’t heard of Claude and has no clue what it’s capable of—OpenAI dominates major partnerships and media attention.

FROM OUR PARTNERS

Close more deals & make more money with this magical AI tool.

At The Neuron, we've already automated big chunks of our sales process using AI.

If you want to do the same and make more money for your business, try Attention ASAP.

It’s an AI sales wizard that:

Listens to your sales calls and tells you what your customers are really thinking.
Automatically updates your CRM based on sales calls.
Sends follow-up emails in under a minute.

It's magic—no joke. Check out their 45-second demo here.

Join 100+ sales teams and start clinching more deals with Attention.

Around the Horn.

Citigroup predicts that AI will automate more banking jobs than any other sector.
Apple has been in talks with several Chinese firms about launching Apple Intelligence in China, where ChatGPT is banned.
YouTube will soon require creators to disclose when their content uses AI; viewers will also be notified when they’re watching AI-generated material.
Forbes sent a letter to Perplexity’s CEO accusing it of stealing content without attribution.

Treats To Try.

Spiral automates 80% of your daily writing tasks while keeping your voice, tone, and style.
Poolside is building an AI coding assistant and just raised $400M+.
Genspark is another AI search engine that answers your questions with summaries (raised $60M).
Daydream is an e-commerce search engine that delivers personalized shopping results (raised $50M).
Materia can answer any accounting questions and also automate parts of an accounting audit (raised $6M).
Day.ai is an AI CRM platform that automates manual data entry by pulling from your convos in email and video meetings (raised $4M).

Intelligent Insights.

Apple’s AI Evolution: Built-in AI, Supercharged Siri, and Ecosystem Lock-in (link).
Andrew Ng On AI Agentic Workflows And Their Potential For Driving AI Progress (link).
Microsoft’s star AI chief peers into OpenAI’s code, highlighting an unusual rivalry (link).
Second Mouse.AI (link).
Microsoft, Nvidia Lead In Investing In AI Startups, But Others Close Behind (link).