When Big Investments Go Wrong

Scale AI's post-investment layoffs expose the messy reality of AI consolidation, OpenAI's Deep Research API tutorial for serious automation, and how to turn any YouTube video into smart docs with Genspark...

In this edition we’ll be covering…

• Scale's workforce cuts right after Meta's massive investment

• A tutorial on OpenAI's Deep Research API

• A tutorial on how to convert YouTube videos into interactive documents

• 5 trending AI signals from Mistral, Anthropic, and more

• 3 more AI tools to add to your arsenal

• And much more…

The Latest in AI

Scale AI Cuts Staff Right After Cashing Meta’s Check

The irony is almost too rich to ignore.

Scale AI just announced they're cutting 14% of their workforce (about 200 employees) barely a month after Meta threw $14.3 billion at them for a 49% stake and poached their CEO Alexandr Wang.

Scale's interim CEO Jason Droege blamed it on ramping up "too quickly" and creating "excessive bureaucracy."

But here's what really happened, their biggest customers started walking away right after the Meta deal…

What's driving the cuts:

• OpenAI has been "winding down" their work with Scale over the past year, cutting ties with their former data labeling partner.

• Google is reportedly also cutting ties with Scale following the Meta deal, according to Reuters, likely because they don't want to fund their competitor's AI ambitions.

• Scale also terminated contracts with 500 global contractors out of thousands, streamlining their own data business operations.

So What?

This is the messy reality of AI consolidation. When Meta writes a $14 billion check, it's a declaration of war. Competitors immediately start cutting ties, creating the exact "market shifts" that force these layoffs.

Scale went from industry darling to cautionary tale in record time. Sure, they're "well-resourced" with Meta's money, but losing your two biggest customers has to be brutal.

Sometimes taking the biggest check means losing everything else…

Tool Spotlight

Master OpenAI’s Deep Research API

OpenAI recently launched their Deep Research API.

Just like in the Web App, instead of just asking questions, this thing plans sub-questions, runs web searches, and synthesizes everything into citation-rich reports. Now, you can do it programatically.

Here's how to get started with serious research automation:

1. Install and authenticate:

pip install --upgrade openai

from openai import OpenAI
client = OpenAI(api_key="YOUR_OPENAI_API_KEY")

2. Choose your model wisely:

• o3-deep-research-2025-06-26 for in-depth, high-quality reports.

• o4-mini-deep-research-2025-06-26 for faster, lighter research tasks.

3. Structure your research request:

response = client.responses.create(
  model="o3-deep-research",
  input=[
    {
      "role": "developer", 
      "content": [{"type": "input_text", "text": system_message}]
    },
    {
      "role": "user",
      "content": [{"type": "input_text", "text": user_query}]
    }
  ],
  reasoning={"summary": "auto"},
  tools=[
    {"type": "web_search_preview"},
    {"type": "code_interpreter", "container": {"type": "auto"}}
  ]
)

4. Extract your research gold cleanly:

print(response.output[-1].content[0].text)

🔥 Pro Tip: Since these tasks take tens of minutes, it’s a good idea to set background=True for complex research. Background mode prevents timeouts while you grab coffee and let the AI do the heavy lifting.

Get Your Hands Dirty!

Vibe Work with Genspark AI Docs

Welcome to vibe working.

Genspark recently made document creation incredibly simple with their AI Docs feature. You can now copy/paste any link or ask it to do research that can instantly generate smart, interactive documents.

Here's a workflow to turn YouTube videos into smart docs:

1. Go to genspark.ai and click on "AI Docs."

2. Copy your YouTube video link and paste it into the prompt window. Say something like "“Create a report for me based on this video: [video link].” You can also choose report templates which is neat.

3. Watch AI Docs summarize, extract key points, or answer detailed questions!

The AI processes the entire video transcript and creates a searchable, interactive document that you can reference and query.

This is perfect for turning long conference talks, tutorials, or lectures into documents, reports, or more!

Quick Bites

Stay updated with our favorite highlights, dive in for a full flavor of the coverage!

Mistral dropped Voxtral, their first open-source audio model that handles 30-minute transcriptions and understands multiple languages.

Anthropic launched Claude for Financial Services, bringing enterprise-grade AI to finance with verified data sources and audit trails.

Grok added anime companions, including a flirty character named "Ani" and an anti-religion panda called "Bad Rudi."

Deloitte rolled out new AI audit tools powered by generative AI to upgrade their auditor capabilities.

Trump announced an AI and energy summit in Pittsburgh at Carnegie Mellon to discuss the intersection of artificial intelligence and energy policy.

Trending Tools

🛠️ Kiro - Amazon's new agentic IDE that goes beyond "vibe coding" with specs and hooks to get prototypes into production systems.

🖥️ ComputerX - AI-powered productivity platform that automates various computer tasks and workflows.

👁️ Copilot Vision - Microsoft's screen-scanning AI that can now see everything on your Windows desktop and help with cross-app workflows.

The Neural Network

Guess “anime companions” are here to stay…

Until we Type Again…

Thank you for reading yet another edition of Digestible AI. Be sure to give us a follow on X, Instagram, and LinkedIn too!

If you have any suggestions or specific feedback, simply reply to this email. Additionally, if you found this insightful, don't hesitate to engage with us on our socials and forward this over to your friends!