Models
Frontier and open models: what shipped, what it can do, and how the claims hold up.
-
GPT-3 shows few-shot learning at scale
A 175-billion-parameter model performs new tasks from examples in the prompt alone.
-
OpenAI withholds the full GPT-2 weights
A staged release holds back the largest model on misuse grounds.
-
BERT makes pretraining the default
Google publishes a bidirectionally pretrained transformer for language understanding.
-
AlphaGo beats Lee Sedol
DeepMind’s system wins four of five games against the Go world champion.
-
Generative adversarial networks introduced
Goodfellow pits a generator against a discriminator, each improving the other.
-
word2vec makes meaning computable
Google publishes a method for learning dense word embeddings at billion-word scale.
-
Reasoning Models Explained: Chain-of-Thought and Test-Time Compute
Reasoning models spend extra computation thinking before answering. The gains on hard problems are real, the costs are substantial, and the tradeoffs…
-
Context Windows Explained: Why AI Forgets, and What Changed
A context window is the model’s entire working memory. Understanding its limits explains most practical AI failures, and why a bigger window…
-
What Is a Large Language Model? How LLMs Actually Work
A large language model learns to predict the next piece of text. That single mechanism explains both what these systems can do…
-
Mistral Large 3
European open-weight option with competitive pricing and a permissive licence.