💶 Free €20,000 if you stay till 2030!!. .
Free AFIRE Guide | AI Academy | Advertise | AI Mastery A-Z
Plus: Most Comprehensive Paid Growth Playbook for Newsletters without Burning Cash
“AI creates better jobs” argument sounds comforting, but Daniel Kokotajlo thinks superintelligence could kill that completely. It’s harder to dismiss than you’d expect.
What’s on FIRE 🔥
IN PARTNERSHIP WITH WORKZONE
Become The Marketer Who Can Build Anything
It’s called vibe coding: you describe what you want in plain English, and AI builds it. Landing pages and campaigns. Dashboards and reports for leadership.
Automations that kill the busywork. No code, no waiting on anyone else. On July 30th, we’ll show you how, starting from the beginning.
AI INSIGHTS
🤖 Why The “AI Will Create New Jobs” Argument Might Be Broken
We have all heard that: “Technology always destroys old jobs, but it creates better new ones.” That was definitely true before. But according to Daniel Kokotajlo, an ex-OpenAI researcher, that historical rule is about to hit a brick wall.
In a recent deep dive on “The Diary Of A CEO,” he laid out exactly why the old patterns of tech adoption do not apply to the dawn of Superintelligence.
-
In the past, machines automated very specific, narrow tasks. Humans simply shifted to the new jobs that emerged, which still required human adaptability.
-
By definition, a superintelligent AI is not a narrow tool. It is actively designed to be better, faster, and cheaper than the best humans at everything.
-
If a brand new job category is created by the AI revolution, a superintelligent system will already be perfectly equipped to do that new job too.
When you zoom out and look at history, human workers always survived massive automation because the human brain was the ultimate fallback plan. We were always more adaptable and capable of general learning than the machines we built.
However, Kokotajlo points out that the entire stated goal of major AI labs right now is to build systems capable of doing all cognitive labor. If they actually succeed in building true superintelligence, the AI doesn’t just replace the “old jobs.”
PRESENTED BY HUBSPOT
Want to get the most out of ChatGPT?
ChatGPT is a superpower if you know how to use it correctly.
Discover how HubSpot’s guide to AI can elevate both your productivity and creativity to get more things done.
Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.
INSIDE AI FIRE NEWSLETTER SYSTEM
💪 We Finally Opened The Paid Growth System Behind AI Fire
For the last 3+ years, we’ve been building AI Fire in public, and along the way, we’ve been lucky to be mentioned by Beehiiv, Paved, Medium, and other people in the newsletter ecosystem.
Recently, many of you told us you wanted more than tips. You wanted courses, live workshops, and even direct help from our team. The newest module covers paid newsletter growth after you already have revenue:
-
Lesson 1 covers the math: CPA, subscriber value, payback, and safe budget levels.
-
Lesson 2 covers the operating system: SparkLoop, Beehiiv Boosts, paid ads, partner reviews, segments, welcome emails, and scaling decisions.
That was the signal for us. We’re also considering a few higher-touch options because many readers asked for more direct help. So before we decide what to build next, we want to hear from you 👇
TODAY IN AI
AI HIGHLIGHTS
😳 Did you see that viral video of Erling Haaland getting scared by his reflection? It was likely AI-generated. Watch both versions and decide which video looks real.
👀 Starting Jul 20 (today), Fable 5 will be limited to Max and Team Premium. Pro and Team Standard users will receive $100 credits for pay-per-use access. More details.
🚨 Kimi K3 demand is sooo high that it just stopped accepting new subscriptions. Existing users are safe, but newcomers now face a waitlist. Try Kimi K3 free here first.
⚡ Qwen3.8-Max-Preview is now available for testing on Qwen Chat and Alibaba Cloud. We still need the benchmarks, but an open-weight release could be next.
🤦 Apple says “nudify” apps are already banned, yet officials still found 8 linked to harmful deepfakes. Three are gone, and several developers could lose their accounts.
🎬 ‘Odyssey’ director Christopher Nolan called AI an obvious ‘Trojan horse’, because everyone knows what is hiding inside. Young people calling out ‘AI slop’ may be right.
💶 Imagine getting €20,000 in ASML shares just for staying until 2030. Around 44,500 employees qualify, as the chip giant prepares for even stronger AI demand.
💰 Big AI Fundraising: Fireworks raised $1.5B at a $17.5B valuation after passing $1B in annualized revenue. Its platform now processes over 40T tokens daily, mostly through customized AI models.
📊 Free Webinar: From Traffic to Pipeline, GTM Lessons from 100 Top AI Products
42DM and Chili Piper analyzed 100 AI infrastructure companies to find out which GTM channels actually turn traffic into pipeline. And you don’t need a data team to apply it.
You’ll walk away with a clear, stage-based framework in one hour. Behind the research:
✅ GTM channels that correlate with revenue at each company stage
✅ Why early-stage SEO often hurts more than it helps
✅ How to build a trust layer before spending on traffic
✅ Website personalization and account identification that convert visits into pipeline
Whether you’re pivoting or scaling fast, this data applies to you. Now it’s time to stop guessing where your GTM budget goes.
NEW EMPOWERED AI TOOLS
-
🔍 OpenSEO gives you affordable keyword research, competitor analysis, backlinks, and site audits to build smarter AI SEO strategies.
-
🛒 Spycost tracks product price history and reveals fake discounts, helping you know when a deal is real before you buy.
-
📈 Kobbe tracks your website traffic, revenue, and funnels with privacy-friendly, cookieless analytics that are simple to set up in seconds.
-
⚡ BaseRT runs local AI models on Apple Silicon up to 6.4× faster than llama.cpp, making on-device AI incredibly fast and easy.
AI BREAKTHROUGH
🔍 NVIDIA Drops New Open Models for RAG and Agent Memory
NVIDIA released a new family of open, commercially available embedding models built to handle agentic retrieval, code searches, and agent memory. If your AI agents often pull up the wrong information, this release is designed to fix that.
-
The flagship NVIDIA Nemotron-3-Embed-8B-BF16 just took first place on the RTEB leaderboard, hitting a 78.5% retrieval score.
-
The RTEB leaderboard measures how reliably a RAG model finds the correct context. When retrieval fails, agents get bad evidence, which leads to extra searches, longer processing times, and higher token bills.
-
NVIDIA also dropped two 1B variants for cost- and latency-sensitive setups. The 1B-BF16 scored 72.4%. There is also a Blackwell-optimized version that can handle up to 2x the traffic while keeping over 99% of the accuracy.
-
All 3 models can process up to 32,000 tokens at once. This makes them perfect for reading long documents, deep codebases, or long agent histories.
To build this family of models, NVIDIA used structured pruning, teacher distillation, and gradually increasing training contexts.
Better retrieval means your agents do less busywork. By feeding the system accurate context on the first try, developers can expect fewer search loops and noticeably lower token costs across the board.
We read your emails, comments, and poll replies daily
Hit reply and say Hello – we’d love to hear from you!
Like what you’re reading? Forward it to friends, and they can sign up here.
Cheers,
The AI Fire Team
Leave a Reply