I had the feeling Haxton had a gun to my head during the whole solving process. "Don't you DARE place an extra glyph or I'll pull the trigger". Bucles is taking cycles and removing all the fun, you can be cooking and realise in the end you don't have money for all the ingredients.

I did not push the metric far as I feared trying risky approaches that would lead to me being 10G above budget and scoring 0 points. Ironicaly I got MGNUP, so I am fighting for tertiary here.

Dual 4 rate pipelines seemed a feasible approach, although using the first input just for air means additional 4 latency. The solution is quasi symmetrical to reduce time spent iterating through the solution space.

I am not confident as I think low 30s are feasible within budget, but it requires a different approach involving at least one 3 rate pipeline or not using the first input just for air. All those attempts were barely above 1k G so probably spiritualshampoo has it. Jammu said they got min+2 but idk what min they were refering too.

Mayybe having C+G/1000 would have been a better approach to budget as it doesnt punish failed attempts that hard. I felt I was fighting against budget and cycles at the same time, CGA for this puzzle would have been as difficult but less painful to do.