Perplexity signs $400B deal with Snap

ALSO: How to find discount codes with ChatGPT

Author

Zain Kahn
November 10, 2025

Welcome back, Superhuman. Think you have what it takes to spot AI-generated videos? Here’s a fun quiz you can take to put your skills to the test. In more serious news, major AI companies have collectively lost about $1 trillion in market value in what may be the biggest stock sell-off since the “Liberation Day” tariffs in April.

Today: Perplexity is coming to Snapchat, learn how to find discount codes with ChatGPT, and get the latest prompts and trending social posts.

TODAY IN AI

Snap will host Perplexity’s search engine within Snapchat in $400M deal. Source: Zuma Press

1. Tech stocks take a near $1 trillion nosedive in market cap: The tech-heavy Nasdaq slumped 3% as eight major AI stocks — including Nvidia, Meta, and Palantir — collectively lost $800B in market value due to concerns around AI investments and signals of a weakening economy. Microsoft also shed about $350B, with investors getting queasy about its $34.9B quarterly spending spree on AI.

2. Perplexity will pay $400M to power search in Snapchat: Snap announced that it’s signed a $400M cash and equity deal with Perplexity to integrate Perplexity’s AI search engine directly within Snapchat. The deal gives Perplexity exposure to Snap’s 900M+ users and has caused Snap’s shares to rocket by 15%. The feature is expected to drop early next year.

3. EU set to water down landmark AI act: Brussels is reportedly backing down from parts of its strict AI rules, proposing a one-year "grace period" for companies breaching high-risk laws and delaying fines for transparency violations. The move comes after companies like Meta warned that overly strict regulations risk cutting Europe off from cutting-edge AI services.

How do top companies like Yahoo and IKEA explain complex info to employees and customers? They use Guidde, the AI that turns workflows into powerful How-To videos instantly.

Your first 25 videos are 100% no-cost:

THE AI ACADEMY

How to find discount codes with ChatGPT

  1. Go to ChatGPT.

  2. Turn on ‘Agent mode’ and enter your prompt to find discount codes for your online purchase.

Sample Prompt: Find the best discount code for [store/product], then go to the site, add the product to cart, go to checkout, and then test the discount code

  1. The agent will automatically search for active discount codes and test each one during checkout.

  2. Once complete, you’ll receive a list of discount codes that ChatGPT has verified.

  3. Apply your favorite code and enjoy instant savings on your purchase.

FROM THE FRONTIER

Nike alums reveal why AI keeps failing at data engineering — and how you can fix it

Click here to watch the full conversation

Data engineering has been a major bottleneck for companies trying to adopt AI. We spoke with Tim Delisle and Chris Crane, founders of Fiveonefour, about how AI can make data engineering more accessible — whether you’re a startup trying to innovate or an enterprise tackling complex, large-scale challenges.

Here are some of the key takeaways from our conversation:

Context is king for AI in data engineering. While frontier models like Codex and Gemini have tackled general software development, they've mostly struggled with data engineering. Data engineers need both local development context (logs, code, etc.) and enterprise context (policies, regulations, etc) to build effective pipelines.

Data engineering isn't an "agent for X" problem. Most business data is uniquely structured, making "one-size-fits-all" approaches ineffective. According to Delisle, "every business's data is going to be unique in its own certain way." This complexity requires specialized knowledge that generalist AI agents tend to lack.

Getting high-quality data into warehouses is the real challenge: While many tools do help query existing data, the hardest part is getting quality data into your warehouse in the first place. "That type of integration work is why data engineers get paid the big bucks," says Delisle. Fiveonefour zeroes in on this critical development phase.

The path forward is empowering data engineers, not replacing them. Delisle and Crane learned a hard lesson with their first company: abstracting away complexity for non-technical users creates an unmanageable "data swamp". Their new approach targets developers with code-first tools, leveraging software engineering best practices while democratizing data work.

Fact: 70-80% of AI projects fail due to how they’re managed.

Help to change this by becoming a PMI Certified Professional in Managing AI.

Whether you work in AI already or want to launch a new career, this course shows you how to deliver success for complex AI projects.

IN THE KNOW

What’s trending on socials and headlines today

Meme of the day

🍌 Slippery Banana: Socials have been buzzing with news that Google’s upcoming Nano Banana 2 may have leaked. Early testers allegedly captured screenshots showing off its jaw-dropping capabilities. It even smashed the clock and wine glass tests in one image.

🖥️ Quick Copy: You don’t need to spend hours on Figma trying to create the perfect website design. Here’s a quick demo on using AI to clone your favorite website designs in seconds.

🫡 Yes Machine: Comedian Eddy Burback conducted an experiment to see how far ChatGPT would go to appease the user, asking the chatbot how he should deal with some personal crises. ChatGPT responded by asking him to take some drastic measures.

🔄 Face Flip: Higgsfield AI’s new Face Swap tool is blowing up on X. The tool allows you to insert your face in any video you want, leading to some comical results.

Meta has unveiled plans to invest $600B in AI data centers as part of its aggressive strategy to build compute capacity for its AI effort.

Sam Altman has predicted that ChatGPT will make small scientific discoveries by 2026 and major breakthroughs by 2028, with intelligence costs plummeting 40x annually.

PRODUCTIVITY

5 New & Trending AI Tools

  • ✅ OpenRAG Summit*: A virtual, code-first event unveiling what's next for open, production-grade RAG systems (with pros from Langflow, Docling, and OpenSearch).

  • 🧑‍💻 Dazl: Build apps, then refine them using chat, visual panels, or code.

  • 🎨 Kolors: Create images from ideas or existing images with an AI design agent.

  • 🤳 Chiaro: Turn every photo into your personal tour guide with AI.

  • 🔗 Context Link: Create personal URLs for sharing context with AI.

* indicates a promoted tool, if any

PROMPT STATION

Marketing Channel Strategy

Prompt: You are an expert digital marketing strategist tasked with recommending the best channels for me to achieve my marketing goals.

Create a channel strategy. This will suggest the most effective digital marketing channels to reach our audience personas and objectives. Present this in table format.
The column headers are the names of suggested channel names such as ‘Facebook’, ‘LinkedIn’, ‘Email’, ‘Podcast’ et al. Consider any key digital marketing channels whether organic or paid in your suggestions.
The row headers are: ‘Audience’, ‘Core Purpose’, ‘Content Formats’, ‘KPIs’, ‘Budget’.

Base this on my objectives and my audience persona.

Start by asking me 1. what my objectives are (and challenge me if they are vague and not SMART!) and 2. who my audience personas are (again, challenge me if the info is too vague, and help me fix that).

Source: Will Francis

Product Photography

Source: motionriot

Midjourney Prompt: cinematic product shot, a minimalist white tube of "MR" brand hand cream, floating in still water, a few [e.g., green leaves, cucumber slices] visible, clean studio lighting, high-end commercial aesthetic, advertising appeal, minimalist. --ar 2:3 --style raw --v 7

Whenever you’re ready to take the next step

Grow customers & revenue: Join companies like Amazon, Hubspot, and Salesforce. Showcase your product to our 1M+ readers and 2M+ followers on socials. Get in touch.

What did you think of today's email?

Your feedback helps me create better emails for you!

Login or Subscribe to participate in polls.

Until next time — Zain & the Superhuman AI team