New research reveals a troubling trend: AI hiring tools show racial bias and systematically reject candidates. What happens then when most employers use the same AI screening algorithms? Read more about the first large-scale study of hiring AI:
RESEARCH
-
Heima Hidden Llama speeds reasoning with abstract thinking tokens
By
–
AI could reason faster by thinking in abstract tokens instead of full sentences! Researchers from Zhejiang University, Adobe, and Northeastern University introduce Heima (Hidden Llama). It compresses lengthy Chain-of-Thought reasoning into a compact set of “thinking tokens,”
-
JudgmentBench: First Public Benchmark for AI Quality Assessment Released
By
–
Congratulations to @StanfordLaw liftlab on the release of JudgmentBench, the first publicly available benchmark in a high-judgment domain where both methods for assessing quality are solicited over the same tasks. Snorkel was proud to contribute as research and data partners on
-
LangChain Labs Announced for Continual Learning in Agents
By
–
ICYMI: We announced LangChain Labs at Interrupt. – An applied research effort focused on continual learning for agents – Early research partners: @NVIDIA
, @harvey
, @PrimeIntellect
, @Fireworks_AI
, @Baseten -
On-Policy Distillation: Emerging AI Post-Training Method
By
–
A new class of post-training method is emerging in 2026: On-Policy Distillation (OPD). It’s already showing up across frontier open-weight model releases, and it’s quickly becoming a technique worth understanding. To help you get up to speed, we’ve compiled a list of the most
-
Opus 4.7 degradation suggests imminent Anthropic model release
By
–
There must be another Anthropic model release coming out soon. Opus 4.7 has been performing noticeably poorly for 2 days now, and temporary model degradation has preceded a new Anthropic model release for several releases now.
-
Recursive Flow Matching: Consistency Across Trajectory Families
By
–
The key distinction: Vanilla flow matching learns a trajectory. Recursive Flow Matching learns consistency across a family of trajectories. That turns step size from a source of error into a training signal.
-
Recursive Flow Matching: New AI Approach for Scientific Simulation
By
–
Scientific simulation has a brutal tradeoff: fast is usually coarse,
accurate is usually expensive. A new paper by Jiahe Huang, Sihan Xu, Sharvaree Vadgama, and Rose Yu proposes a serious way through that bottleneck: Recursive Flow Matching. The target is one of the hardest -
Following Gary Marcus’s 2020 suggestions could have led to AGI
By
–
If we had implemented everything I suggested in my 2020 arXiv paper 'The Next Decade in AI', we might have actually achieved AGI by now. Over the past three years, after a digression motivated by the false promise of simple
-
AI Encoder Tokenizer Performance: 5× Latency Improvement
By
–
At production input lengths, the encoder cuts p50 latency by roughly 5× vs. HuggingFace tokenizers, 2× vs. SentencePiece C++, and 1.5× vs. IREE C. At 514 tokens, it runs in 63 µs with zero heap allocations.