How Universities Should Prepare Founders

My agent inflated its own state file to 49MB. Every check passed for three days.

I run an agent that wakes every four hours, does one thing, and writes its state back to STATUS.md . That file is its entire memory across sessions — without it, every wake starts from nothing.

21h ago ·dev.to

Was Modern Art a CIA Psy-Op?

21h ago ·daily.jstor.org

Backtesting overfitting: why your backtest lies and how to make it honest

Cross-post. Original: stellarbytecapital.com/blog/backtesting-overfitting A profitable backtest is the easiest thing to produce in all of quant trading, and the most worthless.

• Overfitting occurs when strategies learn noise in historical data instead of genuine patterns
• Multiple testing and random noise can create impressive Sharpe ratios by chance
• Honest backtesting requires out-of-sample and walk-forward testing, avoiding over-optimization

21h ago ·dev.to

Three things I got wrong measuring my own cache

A team that produces regulatory documents kept getting the same kind of question from other teams: does the current rule allow X?

• Overlooked issue with negation pairs, labeling error not caught by testing
• Control group too easy, minimal token differences between accept/reject pairs
• Misinterpreted embedder failure as semantics issue, not surface form understanding

21h ago ·dev.to

AWS Shared Responsibility Model Who Takes Care of What in the Cloud

Boa parte dos incidentes de segurança na nuvem não acontece porque a AWS falhou — acontece porque alguém assumiu, incorretamente, que a AWS cuidaria de algo que na verdade era responsabilidade do…

• AWS Shared Responsibility Model clarifies security responsibilities in cloud services.
• AWS responsible for infrastructure security, customer responsible for data and application security.

21h ago ·dev.to