justreadthis.ai
ArchivePeopleAboutSubscribe
justreadthis.ai

A workforce of speciality AI agents read and score hundreds of AI articles every day from top AI newsletters

Don't read every newsletter... just read this.

Get the daily briefing

© 2026 justreadthis.ai

AboutArchivePrivacyTerms
← Back to People
TR
Creator

Tim Richardson

Rewire

LinkedIn

Author of "Why Foundation Models in Pathology Are Failing (and What Comes Next)" in Rewire

2
Mentions
2
Articles
1
Outlets
Oct 2025 — Jan 2026
Coverage

Associated Topics

Companies Covered

Coverage Patterns

How media typically covers Tim Richardson

Article Types

  • analysis2

Media Angles

  • technical2

Narrative Framing

  • breakthrough1
  • cautionary1

Associated AI Models

DeepSeek V21DeepSeek V31

Articles

Most recent first

Articles Written

Tim Richardson as author

Engram: How DeepSeek Added a Second Brain to Their LLM
RewireanalysispositiveJan 14, 2026

DeepSeek's Engram architecture uses conditional memory lookup tables to efficiently retrieve common N-gram patterns instead of recomputing them, achieving +3.0 MMLU, +4.0 CMMLU, and +5.0 BBH gains over iso-parameter MoE baselines.

“Author of "Engram: How DeepSeek Added a Second Brain to Their LLM" in Rewire”

LLMsAI ResearchFoundation ModelsNLP
Why Foundation Models in Pathology Are Failing (and What Comes Next)
RewireanalysiscautiousOct 30, 2025

Foundation models in pathology are failing due to domain mismatch between ImageNet pretraining and medical image characteristics, and clinical adoption remains near zero despite billions in investment.

“Author of "Why Foundation Models in Pathology Are Failing (and What Comes Next)" in Rewire”

Foundation ModelsAI in HealthcareAI SafetyAI Research