If you haven't experienced @Tesla's Full Self-Driving in the past 6 months, it's become mind-bogglingly good.
My wife and I drove up to North Bay this past weekend, and I was so impressed. Over 300 km of watching it make decision after decision. Farm vehicles blocking traffic, no problem. A van pulled out into my lane while we were doing 120km, no problem.
It got me thinking about two new AI models that just launched: Jev from @typesafeai and Laya from Convai Innovations Pvt. Ltd. Both are being called "System 1" models, a nod to Daniel Kahneman's Thinking, Fast and Slow.
Here's why the comparison clicked for me:
Most of the AI we talk about today, like ChatGPT, thinks by writing. It generates an answer one word at a time. Impressive, but slow, and not built for split-second decisions.
Tesla's FSD doesn't write anything. Camera footage goes in, a driving decision comes out, many times per second. Pure reflex.
Jev and Laya bring that same instinct to business software. Instead of generating paragraphs, they take in an email, a support ticket, or a data record and return a decision instantly:
- Which team should handle this?
- How urgent is it?
- How likely is this customer to cancel?
Each answer comes with a confidence score, so the system knows when to act and when to escalate to a human or to the next stage of your automation workflow.
The big difference is that FSD is a specialist. It does one thing, driving, extremely well. I would bet it's better than most human drivers, but that's a spicy topic for a different post ;)
These new System 1 models are generalists. You can hand them a brand new question and get a fast answer without retraining.
Tesla showed us what fast, instinctive AI looks like on the road. We're about to see the same shift inside every business workflow.
Not every problem needs an AI that writes an essay. Sometimes you just need one that makes the right call, fast.
Have you tried FSD lately, or experimented with any of these decision models? Curious what others are experiencing. #AI #ArtificialIntelligence #Tesla #FSD #Automation #FutureOfWork
This weekend I tested Fable 5.1, and it's incredible! 🌟 When Fable 5 came out, I started a brand-new project. Progress went very well, but several times a day I'd get bumped down to Opus 5 due to 'cybersecurity concerns'. I was developing some custom Linux firewall rules, nothing too scary ;) When it happened, I found Opus 5 wayyy too chatty and often drifted more than I liked.
Eventually, Fable 5 felt like it was going in circles, and my progress flatlined. The last PR before this weekend was Aug 14th. Losing three weeks of progress felt like another project that I'd start but never complete. Add it to the pile of software corpses ☠️
Fable 5.1 hasn't disappointed me yet. Not once have I been dropped down to Opus 5. It feels MUCH faster, with the same short, easy-to-understand feedback that came with Fable 5.
Have you tried Fable 5.1 yet? Let me know if it's also much better, worse or the same for your use case.
Are you still on ChatGPT because switching to Claude feels like too much work? This video simplifies the process of how to use ChatGPT memory with Claude AI. I'll show you how to migrate ChatGPT to Claude in three easy steps. #businessleaders
Google just teamed up with Nvidia to allow you to run their new frontier-like Gemma 4 model on local devices.
These models offer advanced reasoning and AI coding capabilities, with variants suitable for mobile devices and powerful GPUs, enabling users to run AI locally.
Your Feed, Your Rules? 🚨TikTok is giving users new power to dial down (or up!) the amount of AI-generated video content they see. This comes as they report over 1.3 BILLION labelled AI videos.
👉 Do you want more control over AI content on your feeds?
#AITransparency#TikTok
Threw this together after having fun with the @AudiCanada RS3 contest last week. It’s 100% AI, created using @higgsfield_ai and Sora 2. Full 4K resolution available on YouTube. Stay tuned for a full breakdown video of how I did it. #audi#audicanada
I had to give it a try ;)
Here's my workflow:
- Source image generated w @ideogram_ai
- Changed the shirt pattern w Nano Banana
- Image upscaled w https://t.co/wDOs4Mdi27
- Video created w Veo 3
- Audio recorded & synced w Davinci Resolve
- All prompts were vetted w ChatGPT 5
Calling all product managers/owners!
If you haven't tried @MiroHQ new Prototype tool, it works well for bringing ideas to life with a single prompt. After the design is created, it is fully editable.
I recommend checking it out now while it's FREE for everyone
Stop being polite to AI. Seriously. Every time you say 'please' and 'thank you', it requires additional GPU cycles, which requires more power. Keep your prompts efficient. #aitips#aitraining#aifundamentals
Is Your AI Lying to You Without Knowing? That's because it's probably hallucinating, which isn't always bad. Here's my quick guide to AI hallucinations 🤖 #aifundamentals#hallucinations#aitips#aitraining