- Everyday AI
- Posts
- The Power And Danger of Letting an AI “See” You
The Power And Danger of Letting an AI “See” You
Oracle unveils AI agents, OpenAI partners with U.S. National Labs, Microsoft adds DeepSeek R1 to Azure AI and more!
Everyday AI
January 30, 2025
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Sup y’all 👋
The past few weeks have been wild, right?
Which one of these below should we tackle in an in-depth episode next week?
After you vote, leave a comment with more details on what you want to see!
What should we cover next week?Vote to see live results |
✌️
Jordan
(Let’s connect on LinkedIn. Tell me you’re from the newsletter)
Today in Everyday AI
6 minute read
🎙 Daily Podcast Episode: AI has become the ultimate resource to grow our companies and careers but have you ever stopped to think about what data is being shared with AI? We break down the pros and cons. Give it a listen.
🕵️♂️ Fresh Finds: GPT-4o gets updates, Gmail AI phishing scam, Microsoft unveils Copilot Surface Pro, and DeepSeek gets restricted by hundreds of companies. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: Oracle unveils AI agents, OpenAI partners with U.S. National Labs and Microsoft adds DeepSeek R1 to Azure AI. For that and more, read on for Byte Sized News.
🚀 AI In 5: Google’s Deep Research is pretty dang good! Is it better than Perplexity? Find out here
🧠 Learn & Leveraging AI: We break down how you can balance getting the most out of AI while keeping your privacy and data safe. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked about DeepSeek hacking OpenAI data, Alibaba challenging DeepSeek, OpenAI safety researcher leaves and DeepSeek under EU watchdog alert. Check it here!
The Power And Danger of Letting an AI “See” You 👁
We're so quick to give AI access to see the world around us, but what are the dangers?
And what are some powers that you're not aware of?
We'll be sharing both as Michael Tiffany, Co-Founder and CEO of Fulcra Dynamics joins us to discuss.
Join the conversation and ask Jordan and Michael questions on AI here.
Also on the pod today:
• Quick emergence of live AI technology 🤳
• Kill switch and intelligent data routing 🎛
• Local compute and requirements 🧑💻️
It’ll be worth your 31 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 – Revfx discovers look-alike accounts instantly, Eapy is a multimedia workspace for music creation and UpMeals is an AI operating system for food businesses.
OpenAI – OpenAI has released new updates to its GPT-4o model.
Google – A new AI-powered phishing scam is targeting Gmail users.
Microsoft – Microsoft has unveiled a new Copilot-focused Surface Pro and laptop PCs.
Trending in AI - DeepSeek’s AI is being restricted by hundreds of companies.
Meta – Meta’s recent earnings topped exceptions and the company is expected to invest more on its AI.
Money in AI - ElevenLabs has raised $180 million in a series C funding, now valued at $3.3B.
Read This – The U.S. Authors Guild is launching a certification for authors to prove their book was not written by AI.
AI Models - Riffusion has released FUZZ, a GenAI music model.
Introducing FUZZ — a generative music model like no other.
Personalized, full-length, high-quality, and infinite. We’re making this instrument free for as long as our GPUs survive.
The best of FUZZ in thread.
— Riffusion (@riffusionai)
3:20 PM • Jan 30, 2025
Ecommerce – Poshmark is adding AI-generated product listings.
AI in Sports – The NFL could replace first-down markers with AI.
1. Oracle's AI Agents Take Center Stage at CloudWorld 👥
Oracle has launched its new AI agents for manufacturers at the CloudWorld event in Austin, aiming to streamline supply chain operations by automating routine tasks. According to Oracle executive Chris Leone, these agents promise smarter decision-making and increased efficiency, potentially revolutionizing how businesses manage procurement and sustainability.
This announcement comes as Oracle joins OpenAI and SoftBank in the ambitious Stargate Project, which includes plans for $500 billion in AI data centers.
2. OpenAI Teams Up with U.S. National Labs 🔬
OpenAI has inked a deal with the U.S. National Laboratories to supercharge scientific research using its cutting-edge reasoning models, according to OpenAI's latest announcement. This partnership aims to thrust U.S. technological leadership forward by accelerating basic science, enhancing cybersecurity, and unlocking new energy resources.
With Microsoft's support, these advanced AI models will be deployed on Venado, an NVIDIA supercomputer at Los Alamos National Laboratory, providing a shared resource for researchers.
3. Microsoft Welcomes DeepSeek R1 to Azure AI Foundry 👀
Microsoft has integrated Chinese AI company DeepSeek's R1 model into its Azure AI Foundry and GitHub. This model, notably cheaper to train than OpenAI's leading alternatives, has sent NVIDIA's market value plummeting by $600 billion.
With the R1 model's cost-efficiency and prowess in AI applications, Microsoft plans to offer a compact version for Copilot Plus PCs soon. Meanwhile, Microsoft and OpenAI are investigating potential unauthorized use of OpenAI’s API in DeepSeek’s development.
4. Microsoft Makes OpenAI’s o1 Model Free to All Copilot Users 🤩
Microsoft is rolling out OpenAI's o1 reasoning model to all Copilot users at no additional cost. Previously exclusive to Copilot Pro subscribers, the Think Deeper feature now allows users to tackle more complex queries by considering questions from multiple angles. With just a tap, users can access step-by-step insights for tasks like app coding or trip planning.
Microsoft AI CEO Mustafa Suleyman expressed his excitement on LinkedIn, highlighting this expansion as a major win for millions of users eager to harness advanced AI capabilities.
5. DeepSeek Raises Privacy Concerns Amid Efficiency Buzz 😳
China's DeepSeek chatbot has stirred privacy worries despite its efficiency. While it's not deemed uniquely dangerous, concerns arise due to China's stringent data laws, which could expose users' data to government surveillance. Experts advise against sharing sensitive information, especially for those at high risk of scrutiny by Chinese authorities.
The advice is clear: proceed with caution and consider alternative ways to engage with AI without compromising personal data.
6. Gemini Levels Up: Google Sheets Gets a Boost 🗒
Google announced on its Workspace blog that Gemini now has enhanced capabilities in Sheets, including answering data-related questions and creating static charts. This update, which is currently rolling out to most Workspace plans and the $19.99-per-month Google One AI Premium plan, promises to save users from the hassle of creating data visualizations manually.
While these AI-generated charts are static and need re-creation with updated data, they offer a quick alternative to traditional methods. However, Google notes that accuracy might be impacted by inconsistent data formatting or vague prompts, underscoring the need for clarity when using this feature.
Google Deep Research: Will this change the way we use the web?
Google recently released Deep Research and it’s REALLY good!
Could it beat out Perplexity? We think so.
We explain why as we break down this update.
Check out today's AI in 5.
🦾How You Can Leverage:
Apparently we just fast-forwarded 10 years into the AI future.
Your memory's broken search function is about to go wild, y’all.
Why?
Because local AI models are turning your morning coffee run into a personal data goldmine.
Let’s be honest.
Your hippocampus is doing its best. But it's running Windows 95 while AI is quantum computing your breakfast order.
Michael Tiffany, co-founder and CEO of Fulcra Dynamics, joined the show today to show us how small AI models running on local hardware are absolutely crushing it at personal data processing.
Like, crushing it HARD.
Forget basic biohacking. We're talking AI that connects your heart rate to your calendar to your location data – then serves up insights your brain could never compute.
This isn't about giving Big Tech more of your precious data. It's about running lightning-fast AI models on YOUR hardware with YOUR rules.
Yeah, we talk about how AI can grow your biz a lot here.
But today, we’re looking at how you can biohack your way to the top with AI, and the pros and cons of letting AI see you every step of the way.
Here’s what you need to know. 👇
1 – Memory Upgrade Has Entered the Chat 🧠
Think your hippocampus is good at location-based memory? That's cute.
Michael discovered something wild about how we store memories. Your brain ties memories to locations – that coffee shop where you had that brilliant idea, that corner where you bumped into an old friend.
His breakthrough? Connect location data to timestamps to meeting transcripts.
Suddenly you've got perfect memory recall. AI doesn't just remember – it connects every dot in your digital life.
Want that conversation from three months ago? The AI checks where you were, pulls up the transcript, then serves the exact moment you're trying to remember. No more "it was somewhere downtown-ish."
Try this:
Start with your calendar and location data only – don't overwhelm yourself with twenty data streams.
Connect them to a custom GPT (Michael did this with zero coding).
Ask specific questions like "Where was I last Tuesday at 2pm and what was I doing?" Once that's humming, add your meeting transcripts from Otter.ai. Create location tags for important convos. Let AI be your memory's search engine.
2 – The Local AI Revolution: In Your Pocket 📱
Eight billion parameters. That's all it takes.
Microsoft just dropped the mic with their latest research. Those massive trillion-parameter AI models? They're shrinking FAST.
Now, there’s SOTA models that can run on consumer hardware.
He's already running small AI models locally for sensitive stuff like home surveillance. No cloud required. The future isn't some massive AI brain in the sky – it's tiny, powerful models running right on your devices.
Your Oura Ring could literally run a decent AI model by next year. Not even kidding.
Try this:
Split your AI stack like Michael does.
Use local models for anything involving cameras or sensitive data.
Start experimenting with edge AI tools that run offline. When privacy matters more than raw processing power, go local.
Save the cloud models for heavy computational lifting. Test different model sizes against your actual use cases.
The smallest model that solves your problem is usually the right one.
3 – Big Tech Doesn’t Want You To Do This 🌊
Your life produces more data than a small business. No cap.
The tech giants want you uploading everything to their clouds. But Michael found a better way.
Build your own personal data lake. Control your own digital destiny.
Every smart device in your house is a sensor. Your car is basically a computer with wheels. Your watch knows more about your health than your doctor.
The power move? Bring it all together on YOUR terms.
Try this:
Start small: Connect your calendar and location data to a custom GPT. Add your health metrics next. Then get spicy – Michael found his smart breakers could track his cooking patterns. What secret data streams are hiding in YOUR smart home?
Build your data lake one stream at a time. Keep it local. Keep it yours.
Big Tech can keep their massive models. You've got something better: YOUR data, YOUR rules, YOUR way.
Now This …
Let us know your thoughts!
Vote to see live results
What are your thoughts? |
Reply