- Everyday AI
- Posts
- Why Artificial Useful Intelligence (AUI) Matters More Than AGI
Why Artificial Useful Intelligence (AUI) Matters More Than AGI
U.S. to reduce AI chip restriction, Apple’s Safari to include OpenAI and Perplexity, U.S. AI leaders testify in court and more!
Everyday AI
May 08, 2025
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Today in Everyday AI
8 minute read
🎙 Daily Podcast Episode: 40 years of AI experience, boiled down to useful advice in less than 40 minutes. Go listen.
🕵️‍♂️ Fresh Finds: OpenAI hires Instacart CEO, Figma adds vive coding AI tool and ChatGPT’s user count is up 25%. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: U.S. to reduce AI chip restriction, Apple’s Safari to include OpenAI and Perplexity and U.S. AI leaders testify in court. Read on for Byte Sized News.
🧠Leverage AI: Can we all stop focusing on AGI and maybe instead focus on AI that makes our lives/jobs better? Let’s learn about Artificial Useful Intelligence, or AUI. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked about Google’s Gemini 2.5 Pro I/O Edition taking the top spot, Amazon working on a new AI code generator and Mistral unveiling a new AI model. Check it here!
Why Artificial Useful Intelligence (AUI) Matters More Than AGI
Maybe we should just skip the whole AGI thing? 🤷‍♂️
And instead focus on something ..... useful?
Ruchir Puri thinks that's the way forward.
Ruchir, IBM Research & IBM Fellow, knows a thing or two about AI and how to make it useful.
For decades, he's helped develop the world's biggest AI breakthroughs, like IBM Watson.
Don't miss this convo if you're ready to make AI a bit more useful.
Also on the pod today:
Why AI is more than IQ đź§
Why chasing AGI might be a waste â›”
How IBM’s leaders are making AI useful 🤖
It’ll be worth your 35 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – Zencoder is an AI coding assistant, Bolt now turns Figma design into apps in one click, and Stori uses agentic AI to build your inbound momentum.
ChatGPT – Sam Altman says that ChatGPT’s user count is up 25% since February.
OpenAI – OpenAI is making Instacart’s CEO its new CEO of applications.
AI Design — Figma is adding a vibe coding AI software design feature.
Google – Google is currently searching for advanced nuclear reactors for AI data centers.
Amazon – Amazon’s new AI tool helps enhance product listings.
AI Healthcare — Concerns have been raised over an AI model that was trained on 57 million NHS medical records.
AI Research – A new study shows that asking chatbots for short answers can increase hallucinations.
1. U.S. to Revamp Biden AI Chip Export Controls ⚙️
In a surprising move just days before a major regulation was set to launch, the U.S. Commerce Department under Trump-era officials announced plans to scrap and simplify the Biden administration’s complex AI chip export restrictions. The original rule, designed to limit China’s access to advanced AI chips by dividing countries into three tiers, is criticized as “overly bureaucratic” and “unenforceable.”
This shake-up could ease global chip access and potentially boost sales for U.S. AI chipmakers like NVIDIA. This change signals a shift toward a more streamlined, government-to-government licensing system, marking a significant moment for the future of AI technology trade and innovation.
2. Apple’s Safari to Shake Up AI Search, Rocking Google’s $20B Deal 👀
Apple’s Eddy Cue revealed last week that Safari searches dipped as users shift toward AI-powered options, signaling a major threat to Google’s dominance and its $20 billion annual payout to stay default on iPhones. Apple plans to add AI-driven search options like OpenAI and Perplexity, challenging Google’s near-monopoly in search advertising and potentially reshaping how billions find information online.
Meanwhile, Google fights back with its own AI features, including an “AI mode” and Gemini AI tech deals, aiming to keep users locked in. This shift could open doors for advertisers and users alike, breaking Google’s grip and ushering in a new era for digital search.
3. U.S. AI Leaders Testify on Staying Ahead in the Global Race 🧑‍⚖️
Top executives from OpenAI, Microsoft, and AMD testified before the Senate Commerce Committee urging increased investment in AI infrastructure and education to maintain U.S. dominance amid rising competition from China’s DeepSeek.
They highlighted the critical need for better export policies and less restrictive regulations to ensure American AI technology remains widely adopted worldwide. With China pushing affordable, powerful AI models, the U.S. faces pressure to innovate faster while balancing national security concerns.
4. Meta’s “Super-Sensing” Smart Glasses Set to Change Wearable AI 👓
Meta is reportedly developing new smart glasses, codenamed Aperol and Bellini, featuring advanced “super-sensing” vision software capable of recognizing people by name, according to The Information. This AI can stay activated with a voice command and track your surroundings to offer reminders like grabbing keys or picking up groceries, promising a hands-free assistant experience.
While current models running live AI drain battery life in just 30 minutes, Meta aims for next-gen devices to sustain hours of use on a single charge.
5. Anthropic Unveils Claude AI with Live Web Search Capability 🔍
Anthropic has just launched a new API that lets its Claude AI models search the web in real-time, making them far more useful for developers needing current data. This update allows Claude to decide when to pull in fresh info, analyze it, and cite sources—no need for developers to build their own search systems.
Compatible with Claude 3.7 Sonnet and newer models, the feature starts at $10 per 1,000 searches and even extends to Claude Code for up-to-date coding help.
6. FDA Teams Up with OpenAI to Speed Drug Approvals đź’Š
The FDA is actively exploring AI to cut down the decade-long drug approval process, with recent meetings involving OpenAI and government AI officers focusing on projects like cderGPT. FDA commissioner Marty Makary highlighted that AI-assisted scientific reviews have just begun, aiming to modernize and accelerate evaluations for treatments, including those for diabetes and cancer.
While no contracts have been signed yet, the agency’s new AI officer Jeremy Walsh is leading these talks, signaling a significant shift toward integrating AI in regulatory science.
🦾How You Can Leverage:
AGI obsession might be wasting our time and money.
Forget the sci-fi future. Focus on what Ruchir Puri calls AUI (Artificial USEFUL Intelligence) instead.
Ruchir is the Chief Scientist at IBM Research and IBM Fellow.
He’s helped build AI for 4 decades — from developing Deep Blue and Watson, Ruchir's seen enough AI hype cycles to know what matters: results, not theoretical capabilities.
He dropped by the Everyday AI show today to cut through the AGI noise and deliver something more useful.
AUI.
Here’s what you need to know.
1. The Triple-Threat Framework That Makes AI Actually Valuable 🥉
Most companies measure AI wrong.
Ruchir told us intelligence isn't just IQ (raw processing power). It's a triple threat: IQ + EQ (emotional intelligence) + RQ (relationship intelligence).
That last one?
Game-changer.
While your competitors chase higher benchmark scores, Ruchir focuses on how seamlessly systems integrate with human workflows.
Because AI that nobody wants to use is just expensive digital furniture.
Try This: Run a 5-minute "Triple-Threat Test" on your AI tools.
Score each 1-5 on: Processing ability, Context understanding, and Human integration. Multiply these scores (max: 125).
Anything below 40? Kill it immediately. Your team will thank you.
2. The 15-Bug Fix That Transformed Developer Life 🪰
Software devs face 40+ bugs daily. Five more "URGENT" ones pop up at 4:55 PM.
Sound familiar? Lolz
Instead of AGI promises, Ruchir's team built something immediately useful: a system that auto-fixes 15 of those 40 bugs.
Not sexy. But life-changing.
And….. useful.
This isn't theoretical value—it's "make-my-kid's-recital" value.
And Ruchir's team measures success by time saved, not future AGI possibilities added.
Try This:
Time-track your three most repetitive tasks for two days. Build tiny AI solutions for just these problems.
Document exactly how many hours you reclaim weekly. These concrete numbers will convince skeptical leaders faster than any AI strategy deck.
Bam. You’ve got more time.
Sounds useful, right?
3. Why Feed-Forward Systems Are Holding You Back 🚸
Ruchir said that current AI is like a rocket launched 0.000001 degrees off course. It'll completely miss the moon.
Ruchir revealed the next breakthrough isn't bigger models—it's self-correcting systems.
Today's "feed-forward" AI needs you to fix bad outputs by changing inputs (endless prompt tweaking).
Tomorrow's "feedback systems" will evaluate their own outputs, compare against your intent, and iterate internally until they nail it.
Simple example: Current AI treats 1,358 Ă— 997 as a language problem.
Future AI will recognize "this is math" and grab a calculator automatically.
AI using the right tools at the right time for the right purposes.
Sounds useful.
Can’t wait for that.
Try This:
Create your "tool inventory" now. List every digital tool your team relies on daily.
For each, note what it does best and when it should be used. This becomes your blueprint for the coming wave of tool-using AI agents.
The teams with organized toolboxes will outperform everyone else.
Get hands-on with AI yourself. Create a strategy for disruption AND reconstruction. And upskill your team before the transition leaves them behind.
The AI revolution isn't in conference keynotes.
It's in the mundane task that just got 30% easier.
That's useful.
Reply