Google's New Gemini API Lets You Trade Speed for Savings
Google just rolled out two new options for developers using its Gemini AI models: 'Flex' and 'Priority.' The Priority tier is for tasks that need top speed and reliability, while the Flex tier is a new, lower-cost option designed for tasks that aren't time-sensitive. It might be a bit slower, but it comes with a significantly smaller price tag. This is a big deal for small businesses using or building custom AI tools. You don't always need an instant answer. For tasks like generating a weekly sales report, summarizing meeting notes after the fact, or creating social media posts in batches, a few extra seconds of processing time is a worthy trade-off for a lower bill. It's about paying for what you actually need. Your move: If you work with a developer or use an app built on Gemini, ask them about switching non-urgent background tasks to the new Flex tier. It’s a simple change that could directly reduce your monthly AI operating costs.
This gives you direct control over your AI spending, letting you pay less for tasks that don't need instant results.
Source: Google AI Blog
Quick Hits
Microsoft Heats Up the AI Model Race
Microsoft just launched three new foundational AI models to compete directly with Google and OpenAI. For you, this isn't just tech news; it's a price war. More top-tier competition means the AI tools you rely on for marketing, sales, and operations will get better features and, most importantly, more competitive pricing.
TechCrunchYour Next Bank Teller Might Be an AI
A company called Gradient Labs is now using OpenAI's models to build AI agents that act as personal account managers for banking customers. This shows where customer service is headed. Your customers' expectations for instant, helpful support are rising. It's a good time to explore simple AI chatbots for your own site to handle common questions.
OpenAI BlogThe Real Reason AI Video Is So Expensive
A new analysis estimates that a single user paying $20/month for an AI video tool can cost the provider a staggering $65 in computing power. This explains the strict usage limits and high prices on these tools. When budgeting for video creation, remember you're paying for raw server horsepower, not just software.
Hacker News AITool of the Day
Apfel
The free, private AI that runs directly on your Mac.
Apfel uses the powerful AI chip already inside modern Macs to run a language model locally. This means your data stays on your machine, it works offline, and it's completely free. It's perfect for quick summaries, rephrasing emails, or brainstorming without sending sensitive info to the cloud.
Install Apfel, then highlight a confusing paragraph in a client email. Press your custom Apfel hotkey and type 'Explain this to me simply.' Get a clear summary in seconds, completely privately.
productivity
Check it out →The Number
$65
That's the estimated monthly computing cost to serve a single AI video user who pays just $20. This massive loss margin shows why AI video tools are pricey and why efficiency is the next big race for AI companies.
Source: aedelon777.substack.com
Know a business owner who needs this? Forward this email.
The best tools are the ones you actually use. Pick one thing from today's issue and give it a try. You'll be surprised how fast small changes add up.
Want this in your inbox every morning?
Join local business owners who start their day with practical AI insights.
Browse All Issues