Bilingual

Wildcard (YC W25) Is Hiring a Founding Engineer
13 days ago
- Wildcard is an agentic commerce optimization platform for ecommerce and retail brands.
- It provides visibility (AEO & GEO), recommendations, execution, attribution, and automation to manage AI shopping agents.
- Brands need insights on product visibility, competitor performance, optimization strategies, and business outcomes as shopping shifts to AI agents.
- The company is growing rapidly at 50% month over month.
- The role is for a Founding Engineer, the first engineer, to shape product, infrastructure, and company direction end-to-end.
- Responsibilities include frontend, backend, queues, browser automation, reliability, customer features, and fast-moving systems.
- Candidates must be high-agency, fast-moving, expert with AI coding tools, and work directly with customers and the founder.
- The recruitment process involves steps like a quick call, system design, a work trial, and offer.
I Cut an AI Agent's Token Use by 94%
8 days ago
- AI agent token use reduced by 94% by compiling stable workflow parts into deterministic code.
- Natural-language skills used initially for flexibility, then specialized harnesses created after analyzing historical traces.
- LLMs only applied for semantic judgment steps like candidate selection and drafting, not for repetitive tasks.
- Optimization pattern: express workflow in natural language, gather traces, compile stable parts, keep LLM calls for key judgments.
- Incentives of big model vendors may not prioritize token reduction, creating opportunities for independent builders.
Tested Kimi K3 for Coding
2 days ago
- The user tested the new Kimi K3 model by asking it to optimize a Rust function, specifically exploring SWAR (SIMD Within a Register) optimization.
- Kimi K3 completed the optimization task perfectly in about 15 minutes, resulting in a 2% speed improvement in the program.
- The user found Kimi K3 to be near-frontier level quality as claimed, costing approximately $1 for tokens on OpenRouter.
METR introduces Expenditure Horizon metric
8 hours ago
- Proposes 'expenditure horizon' as a measure of AI optimization ability, defined as the budget where humans become more cost-effective than AIs.
- Illustrates method using NanoGPT speedrun data: human effort costs ~$2,500 per 1% improvement, while agent runs show expenditure horizons of $0-$3,300 for frontier models.
- Method addresses limitations of binary benchmarks by providing continuous scores and incorporating monetary costs for labor and compute.
- Finds that agent optimization curves are often L-shaped, with agents showing diminishing returns compared to humans, and the approach is most useful for problems with smooth returns to labor.
- Discusses limitations: measures only autonomous optimization, depends on uniform returns, and may be affected by training data contamination or prior agent work.

Hasty Briefsbeta

#ai optimization

Wildcard (YC W25) Is Hiring a Founding Engineer

I Cut an AI Agent's Token Use by 94%

Tested Kimi K3 for Coding

METR introduces Expenditure Horizon metric