Do you love your AI tools - but still spend too much time on rework? You're not the only one. tryfrequency.com/try-frequen…(Works with Claude rn - more to come)

A good barista doesn't need to know your life story. They just need to know exactly how you like your coffee. Frequency does that for your AI: it notices how you like answers and tells your AI, so you stop re-ordering every time. #ai #humanaicollaboration

#humanaicollaboration must be just that. Collaboration. Not just humans bending to adapt to how AI works.

Crazy that pioneering academics and researchers were writing papers in the early /mid 1990s around the possibilities of AI, AI-enabled physical devices, augmented reality, and so on - that are suddenly all becoming reality at once 30 years on

Exploring our first academic partnerships today. Excited to see what we can do together in the #humanaicollaboration space, and what we can learn from real interaction patterns.

If your AI output isn't exactly what you want, first time - there's a reason for that. It's not tuned to how you think. Try our fix to that at tryfrequency.com

Everybody thinks differently. But we all get similar output from AI. What does that mean in practice? If you're using AI as a deep thought partner every day, you're either reshaping your own thinking style to fit that of the LLM you're using, or you're getting frustrated with output that doesn't quite hit the mark. We believe that for AI to be a true thought partner, it should recognize that you don't think the same as everybody else. And it should fit your own patterns and preferences, without resorting to simple sycophancy. Because how you work isn't the same as how someone else works. And you need an AI teammate or teammate that 'gets' your unique thinking style. That's where Frequency comes in, tuning AI output to suit you - not everyone else. #humanaicollaboration

#ai output that doesn’t fit what you wanted costs you time, money – and patience. See our full survey on AI usage patterns: ‘The Average User Doesn’t Exist’ - tryfrequency.com/research-la…

The thing people correct most in AI isn't facts. It's shape. How the output is delivered, more than what is delivered. And that correction costs time and money every time. #ai #humanaicollaboration

Thinking WITH AI is where you do your best work. Yes, AI can do a lot of the thinking, but it's the combination of your brainpower - your judgment, your insights - with AI's that creates the best possible outcomes. #humanaicollaboration

AI should be a thought partner, not just something we hand off cognitive load to. And Frequency tunes it to your wavelength to optimize that partnership. #ai #futureofwork #humanaicollaboration

Offloading isn't the problem. Offloading the wrong things is. #aitools #humanaicollaboration

Research says we often don't really know our own thinking style. So - how you think you think may not be how you really think! #cognitivescience

If humans are involved, the human context matters. #ai #futureofwork #humanaicollaboration

This is Ai tuned to how you actually think and work. Thanks to all our alpha testers so far! #aitools #humanaicollaboration #futureofwork

A yes-man is not a helpful thought partner. #HumanAICollaboration

The re-explaining-the-backstory problem is real and this fixes a big chunk of it. The next layer once that's solved: even with full context, does the agent explain a diff the way this person reads one? Recall and fit end up being separate problems.

Real card from testing this week — it picked up on its own that this person wants the answer before the reasoning. No one told it that; it just noticed. Been staring at these patterns for weeks and they're consistently sharper than I expected.

The part of that new approach that's easy to underrate: agents running long, self-organized work still start every session knowing nothing about how the person they're working for actually wants things done. Capability is compounding. Personalization to the individual hasn't started catching up yet.

Testing Frequency internally today and having to do a lot of it manually to do it properly. Seems old fashioned and laborious but want to do it right. Some things need some good old fashioned graft.