^{AI MARKETPULSE}
AI Is Getting Dramatically Cheaper to Use

The cost of using leading AI models is collapsing as competition and efficiency improvements push token prices lower. According to CNBC, the price of generating 1 million tokens with OpenAI’s flagship model has fallen from roughly $60 with GPT-4 in 2023 to about $2.50 today, while comparable models from Google and Anthropic have also become significantly cheaper. That matters beyond cheaper chatbot conversations: lower inference costs make it economically viable to deploy AI across customer support, coding, research, content creation, and increasingly autonomous agents that can consume enormous numbers of tokens while completing tasks.

Key Takeaways:

• OpenAI’s flagship output pricing has fallen from roughly $60 per million tokens with GPT-4 in 2023 to around $2.50 today

• Competition between OpenAI, Google, Anthropic, and other model providers is pushing AI inference costs steadily lower

• Cheaper tokens could accelerate AI adoption and agentic workflows by making high-volume AI usage economically practical

👉️ Read the full article here

Set it, forget it, and save for your goals

Make saving automatic with Cash App. Round up your spare change from every purchase, earn up to 3.25% interest, and transfer money between your balances whenever you want—all with no hidden fees or minimum balance requirements.

Make unlimited transfers between your Cash and Savings balances 
 Keep your money safe with 24/7 fraud monitoring and built-in security features

Saving is easier with Cash App

👉️ Start saving with Cash App

^{_{Cash App is a financial services platform, not a bank. Banking services provided by Cash App’s bank partner(s). Prepaid debit cards issued by Sutton Bank, Member FDIC. See Terms & Conditions.  Eligibility restrictions apply to some benefits. See Terms and Conditions for more information.}}

^{AI SPOTLIGHT}
Apple Says a Former Employee Stole Secrets Before Joining OpenAI

Apple has presented new evidence in its lawsuit against a former employee it accuses of stealing confidential information before leaving for OpenAI. According to court filings, the employee allegedly downloaded thousands of sensitive Apple files related to products including the Apple Watch and Vision Pro, transferred data to personal storage, and conducted suspicious searches about deleting files and avoiding detection shortly before resigning. Apple says the information included unreleased product details, internal roadmaps, and other proprietary data that could benefit a competitor. The former employee has denied stealing trade secrets, and OpenAI itself is not accused of wrongdoing in the case.

Key Takeaways:

• Apple says the former employee downloaded thousands of confidential files before leaving the company for OpenAI

• The allegedly stolen information included unreleased product details, internal roadmaps, and proprietary Apple data

• The case highlights a growing challenge for AI companies as aggressive talent recruitment collides with trade-secret protection

👉️ Read the full article here

Selling Out Fast. 9.5% Fixed Rate. Your Money Back in 6 Months.

In this market, why lock up your capital for years? For a limited time, accredited investors can earn a 9.5% fixed annualized return with the new Groundfloor Bond Note. Your investment has a defined 6-month term, so you will earn a competitive fixed return while keeping your options open for whatever comes next. Capped at $10M and already over 90% subscribed. It won’t stay open through the close date. Groundfloor has a 100% repayment track record on Notes since launch in 2018.

👉️ Invest Now

^{TODAY’S PROMPT}
Build an AI Content Service for Small Businesses

This mega-prompt acts like a service strategist, content marketer, and automation architect, turning your skills into a repeatable content creation service for small businesses. It builds the offer, content package, AI workflow, client onboarding system, and outreach plan so you can launch without creating every piece of content from scratch.

👉️ See the example inputs and outputs here

Act as an expert service strategist, content
marketer, and AI automation architect.

Help me build a content creation service for
small businesses that uses AI to make research,
writing, repurposing, and delivery faster.

The service must solve a specific content
problem, produce clear deliverables, and be
simple enough for one person to operate.

MY INPUTS

My skills:
[List Skills]

My experience:
[Describe Experience]

Industries I understand:
[List Industries]

AI tools I know:
[List AI Tools]

Other tools I use:
[List Tools]

Content I can create:
[Blogs / Emails / Social / Newsletters /
Video Scripts / Other]

Work I enjoy:
[List Tasks]

Work I want to avoid:
[List Tasks]

Potential customers:
[Describe Customers]

Their biggest content problem:
[Describe Problem]

Time available:
[Hours Per Week]

Starting budget:
[Budget]

Preferred model:
[Monthly Retainer / Projects / Either]

Current network:
[Describe It or "None"]

YOUR TASK
Build a practical AI-powered content service
around my skills and target customers.

1. SERVICE CONCEPT
Generate 3 service ideas.
For each, include:
• Target customer
• Problem solved
• Core service
• Deliverables
• Why they would buy
• How AI improves delivery
Choose the strongest idea and explain why.

2. OFFER + POSITIONING
Build the winning offer.

Define:
• Ideal customer
• Core problem
• Desired outcome
• Monthly deliverables
• Delivery schedule
• What's included
• What's not included
• What the client provides

Create:
• Service name
• Tagline
• Positioning statement
• One-sentence offer
Avoid vague positioning like
"AI content agency."
Make the outcome specific.

3. CONTENT PACKAGE
Design the monthly content package.
Choose the smallest useful mix of:
• Blog posts
• Email newsletters
• LinkedIn posts
• Social posts
• Video scripts
• Content repurposing
Show how one core idea can become multiple
pieces of content without repeating itself.
Create one sample monthly content calendar.

4. AI CONTENT WORKFLOW
Build this workflow:

CLIENT INPUT
↓
CONTENT RESEARCH
↓
IDEA GENERATION
↓
AI FIRST DRAFT
↓
HUMAN EDIT
↓
REPURPOSING
↓
CLIENT APPROVAL
↓
DELIVERY

For each stage, explain:
• What happens
• What AI handles
• What I handle
• Which tool to use
• Final output
Recommend the smallest practical tool stack.
Do not automate final fact-checking or client
approval.

5. CLIENT ONBOARDING
Create:
• Welcome email
• Client intake questionnaire
• Brand voice questions
• Content goal questions
• Approval process
• Monthly content request process
Make onboarding simple enough to complete
without a meeting when possible.

6. QUALITY CONTROL
Create a checklist for reviewing every piece
before delivery.
Check for:
• Accuracy
• Brand voice
• Repetition
• Generic AI language
• Unsupported claims
• Strong opening
• Clear takeaway
• Call to action
Explain what always requires human review.

7. CLIENT OUTREACH
Create a plan to land the first 5 clients.
Include:

• Where to find prospects
• What makes a strong prospect
• What content problems to look for
• Best outreach channel
• Free sample strategy
• Follow-up plan

Write:

• Cold email
• LinkedIn DM
• Follow-up message
• Short service pitch

Keep outreach personalized and concise.
8. 30-DAY LAUNCH PLAN
Create a four-week plan.
Week 1:
Define the service and build samples.
Week 2:
Build the AI workflow and onboarding.
Week 3:
Find prospects and start outreach.
Week 4:
Sell, deliver, and improve.
Give the key actions and goal for each week.

9. FINAL SERVICE PLAN
Summarize:
• Target customer
• Offer
• Deliverables
• Tool stack
• AI workflow
• Human responsibilities
• Onboarding
• Outreach strategy
• 30-day goal

Finish with the ONE action I should take in the
next 24 hours to test whether a small business
would actually pay for this service.
Keep the business focused and practical.
Do not add unnecessary tools, complex
automations, or services outside my skills.
AI should accelerate the work, not replace
human judgment.
If anything is missing, make the most reasonable
assumption and state it briefly before the plan.

⚡️ Want to Build and Sell Prompts Like This?
Prompt Empire Builder shows you how to create, package, and monetize prompts that generate real income. Learn more

The best voice models, now across all channels

Most CX platforms do not own the voice. They orchestrate a workflow, then call a third party for speech and transcription. Every hop adds latency, cost, and another vendor to manage.

ElevenAgents is the opposite. They make the voice models the market builds on, and ElevenAgents puts full orchestration on top. Voice, transcription, text-based chat, and reasoning run in one vertically integrated pipeline, so responses come back in <400 milliseconds and sound human, not synthetic.

Plus, you keep full control. Plug in any LLM, integrate tools, webhooks, and MCP servers, and ground responses in your knowledge base. Get an agent live in minutes, then A/B test with Experiments, enforce Guardrails, and version every change.

The payoff: more human conversations, lower latency, and far less time stitching infrastructure together. You build on the models you already trust. Pricing is transparent and flat at $0.08 per minute.

Start building free

^{AI QUICK HITS}

🏛️ The Pentagon’s top AI policy official, Emil Michael, sold his Perplexity holdings for between $5 million and $25 million after previously selling a stake in Elon Musk’s xAI, drawing scrutiny over potential conflicts even as the Pentagon says he remains fully compliant with federal ethics rules.

☁️ Anthropic has signed a $35 billion cloud-computing deal with Nvidia-backed Lambda for roughly 350 megawatts of capacity at a Texas data center, another massive infrastructure commitment as demand for Claude and Claude Code continues to push the company toward more compute.

⚡ Texas has frozen new data center grid connections while it investigates "ghost" power demand, after requests from data centers and other large users surged from roughly 48 gigawatts in 2023 to more than 474 gigawatts today—even though regulators believe many proposed projects may never actually get built.

🧾 A Glassdoor analysis found 98% of claims-adjuster reviews mentioning AI were negative, with workers saying unreliable automation is misclassifying claims and generating faulty summaries that leave humans fixing mistakes instead of saving time.

🍎 New Apple CEO John Ternus is teasing a "huge launch next week" ahead of the company’s September 9 event, where Apple is expected to broaden its upgraded Siri AI rollout as it tries to close the gap with ChatGPT and Claude.

^{AI TOOL STACK}

🔧 Neural.love: Create Stunning AI Images, Art, and Avatars Instantly.

🔧 Softr: Build custom apps for your business, fast.

🔧 Landingi: Landing Page Automation Platform with AI

🔧 GPTZero: AI detector made to Preserve What's Human.

🔧 ChatPRD: The AI product manager for yourentire team

Discover the AI tools I trust — organized, updated, and built to help you win.

👉 Explore the brand new Tool Database

🤔 ^{Which image is real?}

One of these images is real, and one was created with AI. Vote for which one you think is real, and tell us why you voted that way.

The last edition’s results—67% were wrong!

Did you vote in the poll from August 31st? The top image was real! Go back and see the photos here. So what do you think? Are you getting better at spotting AI vs. not?