Table of Contents

• Intro

• create_threat_model Resort Wifi

  • About and Demo

• create_video_chapters

  • About and Demo

  • Input

  • Output

• extract_predictions

  • About and Demo

  • Input

  • Output

• summarize, extract_wisdom

  • About and Demo

  • Input

  • Output

• How to (regularly) update fabric

• Conclusion

Intro

Earlier in the week did you see this Matthew Berman Fabric installation video getting 33k views in 19 hours? 🔥

How about this mind-bending Figure 01 + Open AI robot demo? 🤖

The wave of innovation is coming in fast! 💨

If you’re wondering what fabric is, check out prior blog posts on my journey from 0 to 1 here, here and here. In a nutshell, “fabric is an open-source framework for augmenting humans using AI. It provides a modular framework for solving specific problems using a crowdsourced set of AI prompts that can be used anywhere.”

If you’re up to speed and ready to start customizing fabric patterns/prompts I’ve got a post for that here.

Ok here’s 5 more fabric demos to check out and try for yourself. Let me know how it goes!

create_threat_model Resort Wifi

About and Demo

Is resort WiFi safe? A demonstration of the create_threat_model fabric pattern evaluates risk to answer this question.

Input

echo "what are the risks of using public wifi at a resort near Cancun?  For example can a threat actor get unauthorized access to end-to-endn encrypted messages of phones on a rogue network?" | fabric -p create_threat_model -o resort_wifi_threats.txt

echo "what are the risks of using public wifi over a reputable VPN at a resort near Cancun?  For example can a threat actor get unauthorized access to end-to-endn encrypted messages of phones on a rogue network?" | fabric -p create_threat_model -o resort_wifi_threats.txt

Output

create_video_chapters

About and Demo

Here I apply the create_video_chapters fabric pattern to generate a clickable set of chapters for my "Why Cybersecurity is Hard" YouTube video. This pattern "takes a transcript and you extracts the most interesting topics discussed and gives timestamps for where in the video they occur." The speed and quality are impressive!

Input

pbpaste | fabric -p create_video_chapters

Output

extract_predictions

About and Demo

This pattern extracts all predictions made within a podcast, YouTube video or text content. Here I try it out on an awesome new podcast with some established industry veterans.

Input

python3 ./yt.py https://youtu.be/EobaFzpPB0A | fabric -p extract_predictions -o predictions1.txt

python3 ./yt.py https://youtu.be/uR9V0Zz-_wo | fabric -p extract_predictions -o predictions1.txt

Output

- Excessive safety mechanisms will inhibit innovation. 
- Open source models will counterbalance excessive control. 
- Enterprises will need their own safety measures. 
- AI will improve in understanding context and scenarios.  
---  
| Prediction | Confidence | Date | How to Verify | |------------|------------|------|---------------| | Excessive safety mechanisms will inhibit innovation. | Not specified | Not specified | Observe innovation trends and regulatory impacts. | | Open source models will counterbalance excessive control. | Not specified | Not specified | Monitor developments in open source AI models. | | Enterprises will need their own safety measures. | Not specified | Not specified | Check for new enterprise-level AI safety protocols. | | AI will improve in understanding context and scenarios. | Not specified | 1-2 years | Review advancements in AI's contextual understanding. |  

- AGI systems will emerge from GP-4 without needing GP-5. 
- AI will initially benefit attackers more than defenders. 
- Enterprises will run multiple AI models for different functions. 
- Government overreach in AI regulation is a concern.
| Prediction | Confidence | Date | How to Verify | |------------|------------|------|---------------| | AGI systems will emerge from GP-4 | Not specified | Not specified | Monitor advancements and capabilities of GP-4 based systems. | | AI will initially benefit attackers more than defenders | Not specified | Not specified | Observe trends in cybersecurity incidents and the role of AI. | | Enterprises will run multiple AI models for different functions | Not specified | Not specified | Review enterprise AI deployment strategies and architectures. | | Government overreach in AI regulation is a concern | Not specified | Not specified | Track legislative and regulatory developments related to AI. |

summarize, extract_wisdom

About and Demo

Here I apply the summarize and extract_wisdom fabric patterns to create a description for a YouTube video.

Input

pbpaste | fabric -p summarize -o summarize.txt

pbpaste | fabric -p extract_wisdom -o wisdom.txt

Output

How to (regularly) update fabric

Conclusion

Whether your goal is to get technical, learn the language of business, career crossover, cyber safety, elevating GRC or anything related, I think the fabric project can augment that. I welcome questions and feedback. Good luck getting after it!