Praveen Vaddadi

out of ignorance, comes certainty - Anonymous

codegen: ai no longer makes things, it makes things up.

Fabs have massive fixed costs but near-zero marginal costs and strong long-term pricing power. AI software is hyper-competitive, supply-constrained today but structurally deflationary. Capital may compound better at the chip layer than in crowded model apps.

In AI, a fab/hardware bet may actually be cheaper than backing yet another software startup.

Santa Claws is a good name for an ecommerce startup of persistent job bots/agents

A timeless gem jstor.org/stable/259121

Pressure masquerades as ambition.

in finance, noise = signal

Understanding isn’t like flipping a light switch; it’s more like slowly turning up the dimmer knob

'estimation' is work.

...and uncalibrated reward/gain scales in the graphs....

Results on standard tasks/tests are relegated to the appendix due to page limits. While the main body seems focused on finding a problem for a specific tool rather than solving a fundamental issue.

Spotting selective evidence in papers (CS): The performance graph shows the new algorithm/data-structure and the baseline tangled together rather than clearly separated.

MCP = stdout as a service.

forkjoin scales because, brent's theorem.