🧰 Picking the most relevant tools one by one leaves gaps. HYSET scores tools as a set. 📊 Recall@5: 88.6% vs. 81.4% ✅ Tasks resolved: 69.9% vs. 61.8% Full story in our weekly newsletter for AI news and research, The Batch 📬 hubs.la/Q04zXp-p0 #DeepLearningAI #AIAgents #AIEngineering

Oct 15, 2026 · 4:00 PM UTC

15
9
49
5,965
Sort replies: Relevant Recent Liked
Replying to @DeepLearningAI
Does set scoring bottleneck retrieval when the tool registry scales?
23
Replying to @DeepLearningAI
so the newsletter scores tools as a set, but does it score the newsletter itself?
38
Replying to @DeepLearningAI
Evaluating tool sets instead of isolated functions just masks poor deterministic state design with exponential token burn. Benchmarks celebrate marginal task gains while completely ignoring the brutal unit economics of context window bloat
1
Replying to @DeepLearningAI
scoring tools as a set instead of one by one makes so much sense. tool selection was always a combinatorial problem, we just pretended it wasn't lol
19
Replying to @DeepLearningAI
that recall jump is actually impressive, nice work.
22
Replying to @DeepLearningAI
Seeing tools evaluated as a set feels like assembling a puzzle where each piece instantly clarifies the picture.
18