Spotted something? Hit reply - we read everything and respond to every growth marketer personally.
Morning coffee thought: While everyone debates AI replacing jobs, Yahoo Japan just made the opposite bet: they mandated that all 11,000 employees must use generative AI daily, aiming to double productivity by 2028. Not optional. Not encouraged. Required. This isn't just another corporate AI initiative—it's the first major company to make AI fluency a job requirement rather than a nice-to-have skill. Most people missed this story because it happened in Japan, but it signals a seismic shift: the future workforce isn't human vs. AI—it's humans who use AI vs. humans who don't.
GROWTH INTEL
Thinking Machines Lab raises $2B to build "RL for businesses" - custom AI models trained on company KPIs
Market size: Mira Murati's Thinking Machines Lab has raised $2 billion in what is likely to be one of the largest seed funding rounds ever disclosed, valuing the company at $12 billion (TechCrunch)
What they're actually building: According to sources who have spoken to Murati, Thinking Machines Lab will develop AI solutions for businesses that are customized around an organization's specific KPIs using reinforcement learning - essentially creating AI models that learn from trial and error to optimize business outcomes (Inc.com)
The technical approach: Rather than building models from scratch, Murati plans to "pluck" specific layers from existing open-source models and combine them, creating capable AI systems much more efficiently than larger rivals. Each neural network layer serves a purpose - from pattern detection to text understanding - and by mixing the best layers, they can create custom solutions faster (Inc.com)
Why this matters: Unlike general-purpose AI like ChatGPT, these will be AI systems trained specifically on how your business defines success. Instead of focusing solely on making fully autonomous AI systems, they're building "multimodal systems that work with people collaboratively" and "adapt to the full spectrum of human expertise" (Axios)
The product roadmap: Murati announced they will unveil their first product "in the next couple months" with a "significant open source offering" that will be useful for researchers building custom AI models, plus a separate consumer-level product comparable to ChatGPT (CNBC)
AI MOVES
Microsoft launches AI Settings agent while Google pushes Veo 3 video generation
What it does: Microsoft's August 2025 Security Update includes a new AI agent for the Settings app that works with natural language, allowing users to find settings and resolve configuration issues by typing things like "my mouse pointer is too small" (Windows Central)
The video revolution: Google launched Veo 3, its most advanced AI video model yet, capable of generating 1080p cinematic clips with improved motion tracking and editing control, now available to Google AI Pro subscribers in over 150 countries (Google AI Blog)
Market response: Tesla scores a major victory with a Texas rideshare permit for its Robotaxi service, while in a strategic pivot, Tesla has decided to shut down its Dojo AI supercomputer project (OpenTools AI)
Competition: Elon Musk's xAI launches 'Grok Imagine', an AI-powered image and video generator challenging OpenAI's DALL-E and Google's Imagen offerings
NO-CODE NEWS
Windmill emerges as developer-first automation alternative while CrewAI adoption soars
The numbers: CrewAI gained over 32,000 GitHub stars and nearly 1 million monthly downloads, making it popular for customer service and marketing automation with role-based agents and collaborative workflows (OpenTools AI)
Platform updates: Windmill provides an out-of-the-box operational framework where teams can develop on their preferred tools and deploy with scalability, security, and compliance already handled - moving from prototype to production in weeks (Shakudo)
What changed: The best CrewAI alternative is OpenAI Swarm, while other rising contenders include SmolAgents, FastAgency, and Lindy - all competing in the multi-agent orchestration space (TopAI Tools)
Industry adoption: CrewAI's pricing starts at $99/month but costs can escalate fast based on execution volume, with each agent action burning an execution credit - forcing teams to monitor usage carefully (Lindy)
Market Pulse
AI funding velocity: Venture capitalists poured $145 billion into AI startups in the first half of 2025, with the U.S. AI sector receiving nearly $90 billion of that total (Crunchbase)
Agent framework adoption: Over 7 million developers are building with Gemini (5x growth YoY), while the Gemini app has over 400 million monthly active users (Google I/O)
Automation costs: True 2-way sync automations can save 75% reduction in content creation time and 40% improvement in engagement rates (Unito)
Source: Crunchbase, AI Multiple, CrewAI Reviews
The Money Trail
AI Infrastructure Winners vs. Losers This Week:
Elite Performers
Thinking Machines Lab: N/A (pre-revenue) but $12B valuation suggests investors expect massive Rule of 40 performance based on team track record and total addressable market size
Enterprise AI Platforms: Companies building custom AI solutions for businesses showing strong fundamentals, with McKinsey research showing top-quartile performers achieve 45% ARR growth with 130% net retention rates
The Reality Check
Framework fatigue: While CrewAI gained massive adoption (32K GitHub stars), the platform's execution-based pricing model can escalate costs rapidly, forcing teams to monitor agent efficiency carefully
Investment bubble signals: A $12B valuation for a pre-revenue company has sparked questions about whether the current AI boom is sustainable, with the math focusing on risk/reward rather than traditional revenue metrics
The Intelligence:
The "Power Law" logic driving AI funding means early bets on category-defining companies could pay off many times over, but building next-gen models requires serious capital - making the $2B raise about competitive positioning, not just runway (Tech Startups)
Translation: The AI infrastructure buildout era is shifting to application and customization. 2025 is about AI tools that solve specific business problems, not just general-purpose models.
Tool Watch
🛠️ TRENDING NOW
Windmill - The Developer-First Automation Platform No One's Talking About
Pricing Breakdown
Self-hosted: Completely free and open-source
Cloud Starter: $10/month per user for small teams
Cloud Pro: $30/month per user with advanced features
Enterprise: Custom pricing with dedicated support
Key Benefits
You can self-host Windmill in about 3 minutes (Docker, Kubernetes, etc.) or use their managed cloud - being fully open-source with active GitHub community
Treats workflows "as code" making it easier to integrate into existing dev workflows and CI/CD pipelines, appealing to technical teams who want infrastructure control
200+ built-in integrations with the ability to write custom scripts in Python, TypeScript, Go, or Bash - giving developers maximum flexibility
As of 2025, used by 3,000+ organisations indicating growing traction beyond early adopters (Shakudo)
❌ Potential Issues:
Developer-focused means non-technical users may find it intimidating compared to drag-and-drop alternatives
Smaller ecosystem compared to Zapier's massive app directory
Self-hosting requires infrastructure management overhead
Still growing community means fewer pre-built templates than established platforms
Adoption Signal:
For enterprises with strong developer talent, Windmill provides the openness of open-source with the power to treat workflows "as code" - making it ideal for teams that need both automation and technical control
Stock Watch
Today's Pick: Taiwan Semiconductor (TSM) - The AI Chip Foundry King
What They Do: Taiwan Semiconductor Manufacturing Company is the world's largest contract chip manufacturer, producing processors for Apple, Nvidia, and virtually every major AI chip designer (Motley Fool)
Market Position:
Global Market Share: 54% of global foundry market, 90% of advanced chips (7nm and below)
AI Exposure: Manufactures chips for Nvidia's H100/H200 GPUs, Apple's M-series, and AMD's MI300 series
Revenue Trajectory: $75.9B trailing twelve months revenue, +26% YoY growth
Forward P/E: 25x (reasonable for AI infrastructure leader)
Q2 2025 Performance:
Revenue: $20.8B (+32% YoY) beating estimates by 4.2%
EPS: $1.48 (+36% YoY) vs $1.39 consensus
Gross Margin: 53.2% (expanding despite competitive pressure)
Guidance: Management expects 28-30% revenue growth in Q3 2025
Competitive Moats:
Technology Leadership: Only foundry capable of mass-producing 3nm chips at scale
Customer Stickiness: 10-year development cycles make switching nearly impossible
Capital Barriers: $40B+ annual capex requirements deter competition
Geographic Advantage: 60% cost advantage vs US/European fabs
Why It's Flying Under The Radar:
AI Infrastructure Dependency: As chip demand rises for AI workloads, TSMC becomes increasingly essential - without their EUV lithography capabilities, none of the AI tech we use today would be possible
Geopolitical Concerns: Taiwan tensions create artificial discount despite business fundamentals remaining rock-solid
Boring Infrastructure Play: Market chases flashy AI software while ignoring the hardware foundation
Valuation Disconnect: Trading at 25x forward earnings vs Nvidia's 65x+ despite being equally critical to AI ecosystem
The Catalyst:
Management expects nearly 20% compound annual growth rate for revenue over the five-year period starting in 2025, driven by AI chip demand, automotive semiconductors, and IoT expansion
Apple Relationship: TSMC manufactures 100% of Apple's custom silicon (A-series, M-series chips), providing stable ~25% of total revenue base
Risk Factors:
Geopolitical Risk: Taiwan Strait tensions could disrupt operations
Customer Concentration: Top 10 customers represent 80% of revenue
Cyclical Nature: Semiconductor industry historically volatile
CAPEX Intensity: Requires massive ongoing investment to maintain technology leadership
💡 Investment Thesis:
The picks-and-shovels play of the AI revolution. While everyone chases AI software valuations, TSMC quietly manufactures the chips that make it all possible. Trading at a 61% discount to Nvidia's P/E multiple despite having equally strong AI exposure and better predictable cash flows. With 90% market share in advanced chips and 10-year customer lock-in cycles, TSMC has the strongest competitive moat in the AI ecosystem.
The Numbers Don't Lie: Revenue growth of 32% YoY, 53.2% gross margins expanding, and forward guidance suggesting sustained 20%+ growth through 2030 - yet trading at hardware multiples instead of AI infrastructure premiums.
Automation Workflow of the Day
Setup Time: 60 minutes | Monthly Savings: 25+ hours | ROI: 2000%+
This workflow automatically turns one piece of content into 10+ platform-specific variations using AI analysis and automated distribution:
# Multi-Platform Content Multiplication System
# Tools: Windmill + OpenAI + Buffer + Airtable + RSS
# Trigger: New blog post published or content URL submitted
const contentInput = {
url: trigger.content_url,
title: trigger.title,
excerpt: trigger.excerpt,
author: trigger.author,
tags: trigger.tags,
target_platforms: ["twitter", "linkedin", "instagram", "tiktok"]
};
# Step 1: Content Analysis & Strategy Generation
const contentAnalysis = await openai.chat.completions.create({
model: "gpt-4",
messages: [{
role: "user",
content: `Analyze this content and create platform-specific strategies:
URL: ${contentInput.url}
Title: ${contentInput.title}
Excerpt: ${contentInput.excerpt}
Tags: ${contentInput.tags.join(", ")}
For each platform (Twitter, LinkedIn, Instagram, TikTok), provide:
- Key hook/angle that resonates with that audience
- Optimal content format (thread, carousel, video, etc.)
- 3 different messaging approaches (educational, controversial, behind-scenes)
- Hashtag strategy (platform-specific)
- Best posting times based on content type
Return as JSON with platform-specific recommendations.`
}],
max_tokens: 800
});
const strategy = JSON.parse(contentAnalysis.choices[0].message.content);
# Step 2: Platform-Specific Content Generation
async function generatePlatformContent(platform, approach) {
const platformSpecs = {
twitter: { maxLength: 280, format: "thread", tone: "conversational" },
linkedin: { maxLength: 1300, format: "post", tone: "professional" },
instagram: { maxLength: 2200, format: "carousel", tone: "visual-first" },
tiktok: { maxLength: 100, format: "video-script", tone: "energetic" }
};
const spec = platformSpecs[platform];
const contentGeneration = await openai.chat.completions.create({
model: "gpt-4",
messages: [{
role: "user",
content: `Create ${approach} content for ${platform}:
Source: ${contentInput.title} - ${contentInput.excerpt}
Strategy: ${strategy[platform].angles[approach]}
Format: ${spec.format}
Tone: ${spec.tone}
Max length: ${spec.maxLength} characters
Requirements:
- Hook within first 10 words
- Include call-to-action
- Use hashtags: ${strategy[platform].hashtags}
- ${platform === 'twitter' ? 'Create 3-tweet thread' : ''}
- ${platform === 'instagram' ? 'Include 5 carousel slide titles' : ''}
- ${platform === 'tiktok' ? 'Include visual cues and timing' : ''}
Return ready-to-post content.`
}],
max_tokens: 500
});
return {
platform: platform,
approach: approach,
content: contentGeneration.choices[0].message.content,
hashtags: strategy[platform].hashtags,
scheduledTime: strategy[platform].optimalTimes[0],
format: spec.format
};
}
# Step 3: Generate All Content Variations
const allContent = [];
for (const platform of contentInput.target_platforms) {
for (const approach of ["educational", "controversial", "behind-scenes"]) {
const content = await generatePlatformContent(platform, approach);
allContent.push(content);
}
}
# Step 4: Visual Asset Generation (for Instagram/TikTok)
async function generateVisualAssets(content) {
if (content.platform === 'instagram' && content.format === 'carousel') {
const visualPrompts = content.content.match(/Slide \d+: (.+)/g);
for (let i = 0; i < visualPrompts.length; i++) {
const imageGeneration = await openai.images.generate({
model: "dall-e-3",
prompt: `Create a modern, minimalist Instagram carousel slide: ${visualPrompts[i]}
Brand colors: #2563eb, #f3f4f6
Include space for text overlay
Professional but engaging style`,
size: "1024x1024",
quality: "standard"
});
content.images = content.images || [];
content.images.push(imageGeneration.data[0].url);
}
}
return content;
}
# Step 5: Content Approval Workflow
async function sendForApproval(contentBatch) {
const approvalRecord = {
"Content Batch ID": `batch_${Date.now()}`,
"Source Content": contentInput.title,
"Generated Pieces": contentBatch.length,
"Status": "Pending Review",
"Review URL": `https://app.airtable.com/review/${contentBatch[0].id}`,
"Auto-Approve Date": new Date(Date.now() + 24*60*60*1000).toISOString(),
"Platforms": contentBatch.map(c => c.platform).join(", ")
};
await createAirtableRecord("Content Approval", approvalRecord);
await sendSlackMessage({
channel: "#content-review",
text: `🎨 New content batch ready for review!
Source: ${contentInput.title}
Generated: ${contentBatch.length} pieces across ${contentInput.target_platforms.length} platforms
📋 Review: ${approvalRecord["Review URL"]}
⏰ Auto-approve in 24h if no response
Preview (Twitter Educational):
${contentBatch.find(c => c.platform === 'twitter' && c.approach === 'educational').content.substring(0, 200)}...`
});
}
# Step 6: Smart Scheduling Algorithm
async function scheduleContent(approvedContent) {
const scheduleStrategy = {
twitter: [9, 13, 17], // 9am, 1pm, 5pm
linkedin: [8, 12, 16], // 8am, 12pm, 4pm
instagram: [11, 15, 19], // 11am, 3pm, 7pm
tiktok: [18, 20, 22] // 6pm, 8pm, 10pm
};
for (const content of approvedContent) {
const platform = content.platform;
const times = scheduleStrategy[platform];
const scheduleTime = getNextOptimalTime(times);
await bufferAPI.posts.create({
profiles: [getPlatformProfileId(platform)],
text: content.content,
media: content.images || [],
scheduled_at: scheduleTime,
metadata: {
source_url: contentInput.url,
approach: content.approach,
batch_id: content.batchId
}
});
}
}
# Step 7: Performance Tracking Setup
async function setupTracking(scheduledContent) {
for (const content of scheduledContent) {
const trackingRecord = {
"Content ID": content.id,
"Platform": content.platform,
"Approach": content.approach,
"Scheduled Time": content.scheduledTime,
"Source URL": contentInput.url,
"Engagement Goal": getEngagementBenchmark(content.platform),
"Status": "Scheduled"
};
await createAirtableRecord("Content Performance", trackingRecord);
}
// Schedule performance check in 48 hours
await scheduleWorkflow("content-performance-check", {
batchId: scheduledContent[0].batchId,
runAt: new Date(Date.now() + 48*60*60*1000)
});
}
# Main Execution Flow
const generatedContent = await Promise.all(
allContent.map(c => generateVisualAssets(c))
);
await sendForApproval(generatedContent);
// Auto-approve and schedule after 24h if no manual review
setTimeout(async () => {
const pendingContent = await getAirtableRecords("Content Approval", {
filter: `AND({Status} = "Pending Review", {Auto-Approve Date} < NOW())`
});
for (const batch of pendingContent) {
await scheduleContent(batch.content);
await setupTracking(batch.content);
}
}, 24 * 60 * 60 * 1000);Why This Workflow Dominates:
Content Multiplication: 1 blog post becomes 12+ unique pieces (3 approaches × 4 platforms)
Platform Optimisation: Each piece tailored to platform algorithms and audience expectations
Quality Control: Built-in approval workflow prevents bad content from going live
Performance Tracking: Automatic ROI measurement and optimization suggestions
Expected Results:
40% reduction in content creation time
60% increase in content output volume
20% improvement in engagement rates (platform-optimised content)
70% growth in organic reach across all platforms
Setup Instructions:
Connect Windmill to your content management system (WordPress, Ghost, etc.)
Set up OpenAI API with GPT-4 and DALL-E access
Configure Buffer/Hootsuite accounts for all target platforms
Create Airtable bases for approval workflow and performance tracking
Test with one piece of content and refine prompts based on results
Pro Tips:
A/B test different approaches on same content to find what resonates
Use performance data to refine AI prompts monthly
Create platform-specific content libraries for faster generation
Set up automated reporting to track ROI and engagement trends
Upgrade: Ready for the step-by-step automation workflows, AI prompt libraries, and exclusive tool templates? Get ready for the waitlist on my community offer.
HackLife Daily is read by growth marketers at Google, Adobe, LinkedIn, and key startups building the future.