从金融研究到蛋白实验：AI 改变的不只是效率，还有工作边界

NLW treats rapid downloads as an encouraging signal, but argues that personal agents still face unresolved retention questions while privacy failures can undermine their claim to dependable utility.

NLW separates accounting charges from operating losses and distinguishes multiyear compute commitments from immediate spending, while stressing that reported valuation targets come from investors rather than confirmed company guidance.

NLW contrasts OpenAI's effort to supply life-sciences tools with Anthropic's reported in-house biology work, while cautioning that these developments do not establish an industry-wide takeover.

Citrini's experiment began as investment research, and NLW says conflicting expert opinions made a bounded cross-disciplinary investigation economically attractive without changing the firm's core identity.

The reported proteins bound to scorpion toxins, but NLW explicitly says neutralization has not been demonstrated, so the result is an early research finding rather than an established treatment.

NLW presents neglected conditions as a potentially high-impact area for AI because lower research costs could make projects viable even when conventional market incentives remain weak.

NLW argues that faster computational work exposes physical verification as a bottleneck, because real-world research still depends on people, institutions, materials, and elapsed time that software cannot eliminate.

NLW expects some competition between AI labs and their customers, but rejects the inference that adding a wet lab will let model providers absorb whole industries at once.

NLW doubts that a scientific breakthrough will automatically repair AI's public image, and argues that labs should prioritize meaningful impact rather than search for a single reputational cure.

NLW calls Citrini's detour an example of opportunity-driven AI, where a firm undertakes previously inaccessible work to improve its core research rather than merely automate existing tasks.