Introducing FrontierRSI, a new recursive self-improvement benchmark by @abundant_labs.
We evaluate agents on (1) a broad range of AI research and engineering work, and (2) post-training models on multi-step agentic tasks, rather than single-step tasks as in PostTrainBench.
https://t.co/7lIBpiCen6
I published my internal "five-whys" skill - it works pretty well for root cause analysis of incidents, outages, etc. We've benchmarked it against smaller prompts and less structured approaches and it justifies its existence.
@Clearpreso It’s all thanks to progressive politics. Including the attitude of the locals. They’re so progressive that voicing unhappiness about the situation gets people publicly dog piled or cancelled
This is from our October internal availability report. We're far from perfect, but we're doing well. The biggest contributor to this is our MySQL databases being hosted on PlanetScale. Also, no sign of our AI acceleration causing serious problems.
Very excited to share that I am returning to @intercom, for one last mission!
I will be working with an amazing team to help improve our customer billing experience.
@karldeeter@cremieuxrecueil What an obscene start to this article.
The 17 year old who was raped by 4 Afghans can console herself that despite their being over represented 5-10x as perps, it’s not an epidemic.
Europe can rediscover its vitality any time it decides to with a coherent culture, debureaucratization, and abundant energy.
Here's a three-step starter pack:
1) Start remigration
2) Drop GDPR
3) Build nuclear
We could start on Monday.