James Tauber
Directory member with public evidence across NLP & language, Compute & infrastructure.
- AI signals
- 1 past 30d
- Sources
- 1 distinct domains
- Discussões
- 21 past 30d
- Latest signal
- 8d ago
Articles & links
sense* Here's a link to the relevant paper cdn.openai.com/research-cov...
Recent commentary
A thought from yesterday’s conference: Philologists are trained to understand culture through texts. Text-based LLMs learn about culture almost entirely through texts. Philology therefore offers a valuable framework for understanding LLMs’ strengths and limitations.
I shouldn’t have expected otherwise but it’s clear many people are getting their information on what the Pope thinks about AI from newspaper headlines, not actually reading (or even skimming) the encyclical.
Did Anthropic clear quotas early to prevent people nearing their quotas from jumping ship to Astra?
I've been saying for a while that one issue with LLM writing is not that it's bad, it's just often a register mismatch. Some software docs it writes sound too much like marketing, for example. Then it just hit me, something like "1,000 Songs in Your Pocket" now just sounds like Claude to me.
Is there prior art for testing and refining annotation guidelines with an LLM even if the annotation will be ultimately be done by humans?
Bad AI takes have nothing on bad SpaceX takes.
In James Tauber's orbit
Center = James Tauber. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.
Are you James Tauber? Show it.
Add the Who’s Who of AI badge to your site or bio. It links back to this profile.
Markdown: [](https://aiweekly.co/whos-who/person/jtauber-com)