Daily Shaarli

All links of one day in a single page.

January 18, 2024

AI: the not-so-good parts - Xe Iaso

As an industry we definitely should think more often about the consequences of our actions. The incentives are indeed pushing us to go faster without much critical thinking.

Willow Specifications

Looks like an interesting protocol for resilient peer to peer data stores. Let's see how it spreads.

Using "will" and "should" in technical writing | James' Coffee Blog
thumbnail

Pick your words wisely. If it does happen every time use "will".

AI poisoning could turn open models into destructive “sleeper agents,” says Anthropic
thumbnail

The tone pointing at "open models" is wrong but the research is interesting. It still proves models can be poisoned (open or not) so traceability and secured supply-chains will become very important when using large language models.

Beware of misleading GPU vs CPU benchmarks
thumbnail

A good reminder that even though GPU tend to be faster, the added complexity and price might not be worth it in the end.

Speedbump: TCP proxy for simulating variable, yet predictable network latency
thumbnail

Looks like an interesting tool to simulate difficult network conditions.