Welcome back to Prompt & Practical, the podcast from the AI For B2B Marketers team where we cut through the noise and get straight to how you can actually use AI in your marketing today.
I’m Ryan, and today we’re covering two big stories: GPT 5 launch timeline and fixes and Demis Hassabis on AI, AGI, and why motivations matter.
It has been a hectic two weeks since GPT-5 arrived. OpenAI shipped, tripped, then sprinted to course-correct. Users were irritated. Limits moved. Model choices reappeared. In this post, I’ll strip out the noise. What actually occurred. What it means for your team. And why I’m paying close attention to Demis Hassabis right now.
Welcome back to Prompt and Practical. I’m Ryan Swindall. I help B2B leaders turn AI into results that show up in pipeline. This week’s topic comes from AI for B2B Marketers.
Here is the sequence. GPT-5 launches. Some users lose the option to select earlier models. Automatic model picking begins working behind the curtain. Rate caps feel tight. Many say output quality feels off. OpenAI replies the next day. Limits rise. Access to other models returns. They acknowledge the auto selector hurt quality. A few days later, controls improve again. You can choose Auto, Fast, or Thinking. Paid tiers get higher ceilings. The model’s tone sounds friendlier.
Why did it feel chaotic? When a tool swaps models without notice, trust drops. Teams run real work on predictable setups. Remove choice on day one and you disrupt production. OpenAI reversed quickly, which helped, but the whiplash was real. Short version: model choice vanished, then returned. Limits were tight, then relaxed. Auto selection underdelivered, then was tuned. The voice felt colder, then warmer.
So what does this mean for teams shipping products and campaigns? Move fast, but narrate faster. If you take options away, explain why and for how long. Keep an escape lane for power users. Migrate later with data, not surprises. Align product, engineering, marketing, and leadership on one storyline told in different words. If your story splits, customers feel it. And when you miss, own it. “We got parts of this wrong” lowers heat and buys space to fix what matters.
If AI powers your work, plan for change. Version your prompts. Keep a clean library with dates, model used, and the job it performs. Test with two providers so you can pivot if one stumbles. Keep a smaller or local model as a backup. For high-stakes work, lock to a specific model instead of Auto. Know your stack. Many SaaS tools sit on someone else’s model. When the base model shifts, your tool shifts. Track that dependency.
Now, Demis Hassabis. AI leaders do not think alike, and that matters. Some optimize for growth and distribution. Ship more. Grab attention. Hassabis speaks like a scientist first. Solve intelligence to accelerate science. Think in decades, not quarters. He keeps a wide range on AGI timing. Maybe this decade, but only if models show steady gains in reasoning, planning, and creativity. Bigger clusters help, but new ideas still matter. We already see models acquire surprising abilities, like video systems inferring physics-like patterns from observation. Motives shape roadmaps. A research-first mindset makes different tradeoffs than a pure growth mindset. One investor-style view: if DeepMind were independent, markets would treat it as a core strategic asset. Alphabet’s future leans heavily on that lab.
Here is the checklist I am using with teams right now. Maintain access to legacy models where possible. Lock critical workflows to a named model and test against a second model monthly. Version prompts and custom GPTs. Tag by model and date. Add a small local model as a safety net. Communicate changes in plain language with screenshots. Measure impact. If quality dips, roll back and explain why.
The headline is simple. Top models are converging. The edge now is orchestration, reliability, and user trust. Teams that win will explain changes clearly, reduce risk for users, and keep a stable lane while pushing forward.
Thank you for listening & subscribing!
🗞️ AI News You Can Use!
5 stories shaping the future of AI and what they mean for your business.
1. AI Now Drives Up to 6% of B2B Organic Traffic—and Growing Fast
AI search isn’t just a sidebar anymore. It already accounts for 2% to 6% of organic visits in B2B—and is accelerating 40% month-over-month. Forrester expects ~20% by year-end. If you're not optimizing for AI-driven discovery, you're leaving a chunk of demand on the table.
2. Agentic AI: Automation That Actually Executes
Forget AI that just offers suggestions. Agentic AI takes the wheel making decisions and taking action on customer interactions and workflows. This isn’t theory it’s a productivity weapon, dramatically reducing admin overhead and closing the gap between insight and outcomes.
3. Deepfake Tech Is Booming, and That’s Risky for Trust
The deepfake market is exploding and forecast to hit billions in growth, but that opens the door to fraud and brand impersonation. B2B marketers, tread carefully. If you're using deepfake or generative content, you better pair it with strong ethics, detection tools, and risk controls.
4. DuckDuckGo lets users block AI-generated images
DuckDuckGo launched a new browser feature that filters out AI-generated images from search results. The toggle, called “AI-generated images,” appears under image search filters and removes content flagged as synthetic — including stock-style renders and AI art. It’s the latest response to growing user backlash against “AI slop”.
5. AI Now Drives Up to 6% of B2B Organic Traffic and Growing Fast
AI search isn’t just a sidebar anymore. It already accounts for 2% to 6% of organic visits in B2B—and is accelerating 40% month-over-month. Forrester expects ~20% by year-end. If you're not optimizing for AI-driven discovery, you're leaving a chunk of demand on the table.
Thanks for reading AI For B2B Marketers! This post is public so feel free to share it with the smartest leader you know.
