Let me put it in a summary for you:
1. R0 value much higher than was in his model
2. Because of 1 and death numbers currently it means there are a lot more asymptomatic carriers than in his model. AKA, less deadly
3. His model came from putting it in his own code that wasn't reviewed by anyone. "I wrote the code (thousands of lines of undocumented C) 13+ years ago to model flu pandemics"
4. His comments about the measures the UK taking improving it has not shown up in data yet. It's conjecture and doesn't account for the 10x plus about face.