If you want a full look-back… Here is the link to my 2023 year in review and 2024 predictions, for those who want to remember what the AI world was like 12 months ago:
RESEARCH
-

What Makes Web 3.0 Unique: Decentralized Internet Revolution
By
–
What makes web 3.0 so unique? #Web3 #web2 #P2P #p2pnetwok #internet #artificalintelligence #AI #blockchain #distributed #semanticweb
-
LLM Training Data Crisis: When Human Content Stops
By
–
You and soooo many others. But if people stop posting/answering questions there, at some point the LLMs will be a dead-end for answers. The LLMs aren't going to train themselves….
-

AI Creating AI: SakanaAI Labs’ First Year Success
By
–
“AI creating AI” 2024 was a blast. @SakanaAILabs has been in operation for just over a year, and we have built not only a world class research team with some of the most creative and innovative minds, but also a world class applied engineering and business team in Tokyo. I feel
-

TLM Tool Reduces Hallucinations in Language Models
By
–
I found a tool to reduce hallucinations in any LLM! TLM (trustworthy language model) is a simple, plug-and-play solution that curbs hallucinations, crucial because LLMs lose real-world value if you can’t trust their outputs. It was tested on OpenAI’s SimpleQA dataset (4,000+
-

Essential AI Benchmarks and Evaluation Metrics for LLMs
By
–
Must Know Benchmarks and Evals: Knowledge: @hendrycks
' MMLU and MATH, @idavidrein
's GPQA and BIG-Bench and their polyunsaturated 2025 variants. Ditto Math lvl 5, AIME, @tamaybes
's FrontierMath, etc Long Context: @ZayneSprague
's MuSR, @realYushiBai
's LongBench, -
Debate over base model benchmarks and training data
By
–
Sure — I’m not disputing it’s an advantage, or even that counterfactually it could have been the only important thing they did. But the benchmarks we see of the base model before any stolen data would be relevant (in post-training) seem to imply otherwise
-

Token-Budget-Aware LLM Reasoning Framework for Efficiency
By
–
Token-Budget-Aware LLM Reasoning The paper proposes a framework for improving the efficiency of reasoning in large language models (LLMs) by dynamically estimating and adjusting token budgets based on problem complexity. This approach helps reduce the token overhead caused by
-

Thousand Brains Project: Neocortex-Based AI Paradigm
By
–
The Thousand Brains Project: A New Paradigm for Sensorimotor Intelligence The Thousand Brains Project aims to create AI systems based on the principles of the neocortex, focusing on sensorimotor learning and interaction with the environment to build world models, similar to