- AI Fire
- Posts
- 🏛 TradeGPT Trump Flexed $10B
In partnership with
Read time: 5 minutes
Trump posted an AI image of himself day trading Intel stock and it weirdly matched a real $10B move by the US. Also, xAI just dropped Grok 4 Fast, and it’s shockingly good
What are on FIRE 🔥
IN PARTNERSHIP WITH BELAY
Economic pressure is rising, and doing more with less has become the new reality. But surviving a downturn isn’t about stretching yourself thinner; it’s about protecting what matters most.
BELAY matches leaders with fractional, cost-effective support — exceptional Executive Assistants, Accounting Professionals, and Marketing Assistants — tailored to your unique needs. When you're buried in low-level tasks, you lose the focus, energy, and strategy it takes to lead through challenging times.
BELAY helps you stay ready for whatever comes next.
AI INSIGHTS
xAI released Grok 4 Fast, a lightweight, cost-efficient version of their full Grok 4 model and it immediately landed in the top 10 global models! It runs ~25× cheaper than Gemini 2.5 Pro and even outperforms Grok 4 on some tasks.
It’s uses ~40% fewer tokens per prompt & also runs in 2 modes:
“Thinking mode” for complex reasoning
“Fast mode” for quick answers
…all in one unified architecture. Let’s just say it punches above its weight class:
#1 in web search performance (LMArena Search Arena)
#8 globally for prompt intelligence (Text Arena)
#1 on LiveCodeBench, beating GPT‑5 and even Grok 4
92% on AIME 2025, 93.3% on HMMT → math scores that rival GPT‑5
And here’s what’s really wild, it delivers all that while using 61M tokens to finish benchmark suites vs. 93M for Gemini 2.5 Pro and 120M for Grok 4.
Why it matters: You’re basically getting Claude+Gemini level power for indie-level pricing. It’s for devs, agents, and products that care about function over flair (A direct shot at the “OpenAI pricing wall”)
PRESENTED BY GURU
Become an email marketing GURU.
Join us for the world’s largest FREE & VIRTUAL email marketing conference.
Two full days of email marketing tips & trends, famous keynote speakers (Nicole Kidman!), DJ’s, dance contests & networking opportunities.
Here are the details:
100% Free
25,000+ Marketers
November 6th & 7th
Don’t miss out! Spots are LIMITED!
TODAY IN AI
AI HIGHLIGHTS
🎓 Oxford just gave every student and staff member Free access to ChatGPT Edu, the first UK university to do it. If you're here, congrats, check out to see your access here.
🧾 Anyone knows ChatGPT. But there’re some AIs you've never heard of. So reporter Alex tested all of them. Some, impressive. Here's how they stack up against ChatGPT.
✈️ A traveler just shared how he booked a $879 flight for $299 with no points or memberships. Here are the 8 prompts he used to make ChatGPT his travel agent.
📈 Trump memed himself trading Intel stock… but it’s really a flex about a $10B government bet on American chip dominance. Totally real money. Here's the image.
🛡️ LinkedIn will soon share your profile + activity data with Microsoft for AI training and ad targeting, starting Nov 3. Default settings are ON, so opt out if you care.
💰 AI Daily Fundraising: Israel’s Glilot Capital just raised $500M to back AI and cybersecurity startups. Over $700M already flowed into their portfolio this year.
AI SOURCES FROM AI FIRE
NEW EMPOWERED AI TOOLS
🍌 Nano Banana Biz clones & customizes your own Nano Banana
📚 TutorFlow turns your idea into a course with a single prompt
⚡ Ask AI For It gets 15000+ AI tools to boost your productivity
✍️ Paraphrase Tool rewrites, humanizes AI & creates social bios
AI QUICK HITS
👓 Meta is letting outside devs build apps for its smart glasses
⚙️ X is apparently moving to a customizable AI-powered algorithm
🎬 Big tech is under pressure from new AIs. Will this change?
🚨 Chatbots depicting child abuse images spark AI misuse fears
💼 CEOs are so bullish on AI that they’re ignoring everything else
AI CHART
Forget easy benchmarks. SWE-Bench Pro just raised the bar and even top AI models are struggling. If you’re used to AI benchmarks where models score 70–90%, this is not that.
In short: if your model passes, it’s actually solving the problem. So… how did the top models do? It’s a bloodbath on SWE-Bench Pro:
OpenAI GPT‑5 scored just 23.3%, down from ~70% on SWE-Bench Verified.
Claude Opus 4.1 hit 23.1%, also falling from ~70% previously.
On private commercial code, GPT‑5 sank further to 14.9%, and Claude to 17.8%.
GPT‑4o? Just 4.9% on public tasks.
And yet, top models still do better than the rest, and their performance is more consistent across languages and repos. Until then… maybe don’t fire your engineers just yet.
This benchmark is the realest test we’ve seen yet. SWE-Bench Pro gives us the map. And the gap. Now we know exactly how far agents need to go before they’re ready for production, and which models are closest.
PRESENTED BY ROKU
It’s go-time for holiday campaigns
Roku Ads Manager makes it easy to extend your Q4 campaign to performance CTV.
You can:
Easily launch self-serve CTV ads
Repurpose your social content for TV
Drive purchases directly on-screen with shoppable ads
A/B test to discover your most effective offers
The holidays only come once a year. Get started now with a $500 ad credit when you spend your first $500 today with code: ROKUADS500. Terms apply.
We read your emails, comments, and poll replies daily
How would you rate today’s newsletter?Your feedback helps us create the best newsletter possible |
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
Reply