Steal an expert’s brain in 20 mins

👋 What’s up friends!

This is Prompt Punk — the best AI sales newsletter in the Infiniverse! 😉

A little shorter this week, as I have a sick kid at home. Less computer time.

TL;DR: inside today’s newsletter

Let’s rip…

The other day I was watching a YouTube guru explain how to grow an email newsletter.

It’s common for me to run to YouTube when I have a problem. Like assembling a Bugaboo stroller.

Days later, I wish I could ask the newsletter guru some specific questions about my situation, with my data. But I didn’t want to send a cold email and wait days.

So, I used AI to create an on-demand advisor based on the guru’s public content.

Here’s what I did…

Result: within 20 minutes I had clear answers to my questions and a concrete plan tailored to my newsletter.

Obviously, you can do this for anything. Sales advice, dating advice, getting a promotion, carpentry, whatever.

Why this works

Do the work once, and you’ll have an on-demand expert you can consult whenever you need a decision.

A side-by-side clip circulating this week shows just how far generative video has come: early-2023 ModelScope on the left vs. xAI’s new Grok Imagine 0.9 on the right. The difference in facial coherence, motion, lighting, and overall realism is night-and-day.

Press play…

Deloitte will make Anthropic’s Claude available to 470,000 employees. Anthropic’s largest enterprise deployment to date.

The alliance includes a Claude Center of Excellence to move pilots into production and a certification program for 15,000 Deloitte professionals.

How they win
Deloitte gets a unified AI stack and a trained delivery army; Anthropic lands flagship distribution inside a Big Four.

Why it matters
This is scale, not experiments. When a firm serving ~90% of the Fortune Global 500 bakes a model into daily workflows, AI moves from “innovation” to operating procedure and Claude becomes the reference choice.

Expect faster enterprise rollouts, more certified practitioners on client projects, and copycat deals from rival consultancies.

Google pulled its rezoning bid for a 468-acre datacenter campus in Franklin, Indiana minutes before a council vote widely expected to reject it. Local coverage and industry press peg the project at $1–$1.5B in capex.

How they lose

Google: Delayed time-to-compute, higher friction on future siting, and reputational drag with communities worried about water, power, noise, and light. Especially when jobs are relatively few versus incentives. The withdrawal followed months of organizing and a packed council meeting with overflow rooms.

Indiana: Forfeits a billion-dollar-plus build, construction spend, and long-tail tax base tied to ancillary development, while signaling to other hyperscalers that approvals may be unpredictable. (Council opposition centered on abatements and utility burdens.)

Why it matters
AI data centers are scaling faster than local permitting comfort. Expect more flashpoints where power, water, and incentives collide with jobs, noise, and land use and more last-minute exits if votes look hostile.

What it does: End-to-end outbound: find B2B contacts (450M+ database), run multi-inbox campaigns with rotation and deliverability safeguards, personalize with AI, and manage replies in a unified inbox + lightweight CRM.

Example: A customer reported: 40+ meetings in 9 days from a single campaign (38.3k sends, 2.2% reply rate). Instantly’s playbooks emphasize tracking replies/positives/meetings and staying under Gmail bulk-sender thresholds.

Why it’s valuable: Consolidates list-building, testing, email sending and compliance into one stack. Fewer tools to juggle, faster A/B cycles to get your meetings booked.

Website: instantly.ai

What it does: Automates the grunt work editing long-form videos. Finds the key moments, edits them with captions/branding, and publishes straight to TikTok, YouTube, and Instagram. Think “podcast/webinar → social clips,” with agents that can search and analyze your whole library.

Example: Drop in a 60-minute show. Overlap finds 8–12 coherent, standalone clips and auto-posts them on schedule.

Why it’s valuable: Turns existing content into a steady pipeline of on-message short video. More coverage, less manual editing.

Website: overlap.ai

xAI is raising $20B to convert cash into compute, closing the gap with OpenAI and Anthropic on raw capacity.

The strategy is straightforward: dominate capacity, shorten training cycles, ship better Grok models faster. xAI’s bet is speed to power and GPUs, compressing build timelines from the usual years to months, keeping Grok’s AI benchmark performance moving.

OpenAI’s 6-GW datacenter pipeline suggests eventual top-end capacity leadership; Anthropic’s AWS + Google setup aims for scale with diversified silicon. In short: xAI = build fast, OpenAI = buy huge, Anthropic = scale with cloud partners. All three are racing to convert power into model quality and release pace.

Why it matters:
If Musk can keep compressing build timelines and stacking power/GPUs at gigawatt scale, Grok’s improvement cadence becomes a function of infrastructure.

Musk is aiming to get to Artificial General Intelligence first, while others mess around with consumer apps, enterprise sales and other “side quests”.

This week OpenAI added “apps” to chatGPT. It’s a continual play to become a platform, like Microsoft Windows in the 90’s.

OpenAI’s new “apps inside ChatGPT” and Apps SDK lets third-party services run natively in the chat window. In demos, Canva generated and refined designs inline, and Zillow rendered an interactive home search map, no handoff to a separate site or app.

Strategically, this shifts ChatGPT from an “assistant” to a distribution platform: users stay in one place while developers bring their workflows to it.

Why it matters:
Placement inside ChatGPT becomes the modern shelf space (like Google ranking). If your brand wins the default slot for a high-value intent, you capture demand the moment it appears, with no tab-hopping or drop-offs.

Google DeepMind released Gemini 2.5 Computer Use. An AI agent that can see a live screen or browser and propose exact UI actions (clicks, keystrokes) with lower latency. It lands in a landscape already seeded by OpenAI’s Operator and Anthropic’s own Computer Use. You can expect it to plug neatly into workflow stacks like n8n where ops teams orchestrate these tasks across CRMs, billing, and the web. I’ll even have a look at it for the AI sales agent I am building.

Why it matters:
We’ve seen the success of coding agents, this is the birth of ops agents. They point straight at IT ops, sales ops, finance, and support.

All planets revolve around the sun — NVIDIA.

Thanks for reading! Peace and Love! ❤️

If you liked this:

📨 Forward to a biz techie friend
💬 Hit reply and tell me what you want more (or less) of
📊Vote in the poll below 👇

— John
Prompt Punk