Turns out AI is RACIST
Plus, 🧠 How to use multiple AI models together for better research and writing, Huawei and DeepSeek both expanded China’s push toward technology self-sufficiency recently, and more!
Hola Decoder😎
If someone forwarded this to you and you want to Decode the power of AI and be limitless, then subscribe now and Join Decode alongside 30k+ code-breakers untangling AI.
🤖 A Study of 4 Million Job Applications Just Found AI Hiring Tools Are Racially Biased
A Stanford-led study analyzing 4 million applications across 156 major employers found clear racial disparities in AI hiring algorithms, while a separate Financial Times report reveals AI is simultaneously reshaping how top consulting firms charge for work.
The Decode:
AI Hiring Tools Are Discriminating at Scale
The study found 10.62% of positions showed adverse impact against Black applicants and 5.32% against Asians, with over 1 in 4 Black applications landing at federally flagged roles.
42 models are shared across employers, so a rejection at one company follows you to another running the same algorithm. 4% of applicants who applied to 10 roles were rejected from every single one.
AI Is Dismantling Hourly Consulting Fees
According to the Financial Times, clients no longer want hourly billing when AI handles analytical work in minutes. McKinsey is already restructuring partner pay to absorb the revenue unpredictability.
Law firms and auditors are facing the same pressure, with clients across industries demanding fees tied to actual outcomes rather than time spent.
These findings cover only Pymetrics and may not apply to modern AI hiring tools that work differently. But when one vendor dominates an industry, their blind spots become everyone's problem.
Together with Stack Influence
This Channel Delivers 5X ROAS and Gets You to Page One in 60 Days
Ancient Nutrition had a real product. Strong formulation. Decent listing. What they didn't have was sales velocity that the algorithm could work with. So they stopped throwing money at bids and started sending product to micro-influencers instead.
The result: 15x sales growth, a 7.2x rank boost, all from creators their customers actually trusted.
Stack Influence automates the motion. Their AI matches your brand with creators from a network of 11M+ micro-influencers, and you simply pay them only with your products, so every dollar maps to a real purchase and a real piece of content.
- Neutralyze hit $233K in revenue at a 3.8x ROI on product seeding alone
- Lenny & Larry's activated 1,560 creators and hit 11x monthly sales
- Aunt Fannie's gave free product to 189 creators, and 8x'd Amazon sales in 90 days
Now you can run the same motion, and save 10% in your first campaign if you sign up this week.
🧠 How to Use Multiple AI Models Together for Better Research and Writing
Different AI models are good at different things. Claude might write better, ChatGPT might structure ideas better, and Gemini might surface different insights entirely. Anuma lets you use all of them together in one workflow without losing memory or repeating context.
Step 1: Open Council Mode - Enter one prompt once, then send it to ChatGPT, Claude, Gemini, Grok, DeepSeek, and Kimi simultaneously from the same screen.
Step 2: Compare responses side by side - Use Claude for deeper writing, ChatGPT for structure, Gemini for broader ideas, and Grok for faster takes without switching tabs or copying prompts between apps.
Step 3: Continue the conversation with any model - Pick the best response and continue refining it while your context, preferences, goals, and chat history stay shared across every AI model.
Step 4: Build a persistent AI workflow - Use Anuma through desktop, iMessage, or SMS so your AI assistant remembers ongoing projects, writing style, and previous conversations automatically.
Everything stays encrypted locally on your device with no logs, no training on your data, and full export control, turning multiple disconnected AI tools into one continuous workspace. You can try Anuma here.
🇨🇳 China’s AI and chip race keeps accelerating
Huawei and DeepSeek announced aggressive new moves across semiconductors and AI pricing, showing how Chinese companies are pushing harder toward domestic infrastructure independence and lower-cost competition against U.S. tech leaders.
The Decode:
Chinese AI competition intensifies - DeepSeek permanently reduced pricing for its flagship V4 Pro model by roughly 75%, making it significantly cheaper while increasing pricing pressure across the frontier AI market.
Anthropic criticism resurfaces again - The pricing move may reignite earlier tensions with Anthropic, which previously accused DeepSeek of using Claude-generated outputs to help improve and train its own AI systems.
China pushes semiconductor independence - Huawei said it has created a practical pathway toward producing chips competitive with TSMC’s next-generation 1.4nm process by 2031 despite years of restrictions from U.S. semiconductor sanctions.
China shifts away from U.S. chips - The announcement comes as Chinese authorities continue restricting domestic companies from purchasing Nvidia and other American chips, increasing pressure to build competitive local alternatives.
China’s AI strategy is increasingly focused on two things simultaneously: building domestic infrastructure and aggressively competing on cost. The broader pressure on U.S. companies may come not only from capability advances, but from pricing and supply chain independence.
Together with Playbook Pro
Your Growth Strategy Is Costing You. Here’s the Math
Ad costs are climbing. Most first-time buyers disappear within months, and margins are shrinking fast when discounts drive sales.
We spent 200+ hours on research, distilling what top brands do into a field guide you can read and implement in under 30 minutes.
Inside Rewriting the DTC Rulebook, you’ll learn how to:
📈 Lift retention by 5% to unlock 25–95% profit growth
🔥 Launch emotional campaigns proven to double repeat sales in 8 weeks
💰 Use the Brand Value Canvas to raise prices 15–40% without losing volume
🧠 Slash decision cycles from 3 weeks to 3 hours using the Assumption Buster Toolkit
📊 Cut paid-ad dependence by 30% in 60 days with the Owned Media Checklist
🏆 Tools you Cannot Miss:
💡 IdeaPico – Guides you through structured questions to turn rough ideas into detailed app briefs ready for AI coding agents.
🧠 MemoryBase – Saves and organizes your ChatGPT and Claude conversations into a searchable memory layer you can reuse across AI tools.
🎨 ImgLume – Generates polished AI images from simple text prompts for creative and marketing use cases.
🎵 Melodio – Creates fully produced songs with AI-generated lyrics, vocals, and instrumentation based on any mood or story prompt.
📈 FN2 AI – Monitors your portfolio, earnings, and market activity with AI agents that deliver concise investor briefings automatically.
🚀 Quick Hits
🔍 Backlash to Google’s AI-first Search redesign appears to be boosting rivals, with DuckDuckGo reporting a 33% jump in U.S. iOS installs and up to 27.7% higher usage of its “No AI” search mode.
👋 Dropbox co-founder Drew Houston is stepping down after building the company into a globally used cloud storage platform, and says his next entrepreneurial focus will be AI.
💰 AI gateway startup OpenRouter raised a $113 million Series B led by Alphabet’s CapitalG, reportedly doubling its valuation to $1.3 billion as demand surges for multi-model AI infrastructure.
🎵 Universal Music Group and TikTok renewed their licensing deal with a stronger focus on AI safeguards, including commitments to remove unauthorized AI-generated music and improve artist attribution on the platform.
🎶 ElevenLabs launched Music v2 with improved vocals, instrumentation, arrangements, and multilingual generation, powering its ElevenMusic creator platform, upcoming music generation API, and downloadable soundtrack tools for ads and branded content.
🧩 Prompt of the Day
Frequency Capping That Protects Ad Performance
Most campaigns don’t fail because audiences never see the ad. They fail because audiences see it too many times and stop paying attention.
Control repetition before fatigue kills performance.
Paste the prompt: Drop this into ChatGPT, then fill in your campaign details.
Prompt to paste
Create a frequency capping strategy for [Insert campaign or brand] across platforms such as [Insert platforms]. Recommend the ideal maximum impressions per user within a defined timeframe and explain how frequency impacts CTR, engagement, conversions, and ad fatigue. Suggest different caps for prospecting, retargeting, and awareness campaigns, and include monitoring guidelines for identifying when creative fatigue begins. Focus on maintaining ad effectiveness while preventing audience burnout.
🤳AI Nugget of the Day
Thanks for Decoding with us🥳
Your feedback is the key to our code! Help us elevate your Decode experience by hitting reply and sharing your input on our content and style.
Keep deciphering the AI enigma, and we'll be back with more coded mysteries unraveled just for you!