The AI terms that
actually matter.
Clear, technical definitions of 146 key concepts — written by AI engineers, not marketers.
HyDE (Hypothetical Document Embeddings)
HyDE is a zero-shot retrieval technique that uses an LLM to generate a hypothetical document based on a query, and uses that generated document's embedding to find real relevant documents.
Read definitionInflect v2
Inflect v2 is a pair of open-weight, VITS-family English text-to-speech models, Inflect-Micro-v2 (9.36M parameters) and Inflect-Nano-v2 (3.97M parameters), that generate complete 24kHz waveforms locally on CPU with no external vocoder or cloud dependency.
Read definitionJailbreaking
Jailbreaking is the practice of crafting adversarial prompts or multi-turn conversations that manipulate an LLM into bypassing its safety training and producing content or actions it was explicitly aligned to refuse, remaining one of the most persistent unsolved problems in AI safety despite continuous defensive advances.
Read definitionJev
Jev is TypeSafe AI's 'System One' model, released in early access on September 15, 2026: instead of generating text, it takes a block of state plus a set of typed questions (Choice, Score, or Noul) and returns calibrated probabilities and confidence scores in one parallel pass, positioned as a faster, cheaper substitute for using a full LLM to make repeated structured decisions.
Read definitionKimi K2.7 Code
A frontier scale, coding focused agentic AI model by Moonshot AI. It features a 1 trillion parameter MoE architecture with advanced multimodality and long horizon software engineering capabilities.
Read definitionKimi K3
Kimi K3 is Moonshot AI's 2.8 trillion parameter open-weight Mixture of Experts model, the first open model in the 3-trillion-parameter class. It combines a hybrid linear/full attention mechanism (Kimi Delta Attention), an extremely sparse 16-of-896 expert MoE layer, and a native 1M-token multimodal context window.
Read definitionKling 3.0 Model Series (Omni)
Understand professional AI video and image creation with Kling 3.0. Explore Video 3.0 Omni’s multi-shot cinematic control and Image O1’s precision editing for high-fidelity, agentic workflows.
Read definitionKnowledge Distillation
Knowledge distillation is a model compression technique where a compact 'student' model is trained to replicate the behavior of a larger, more capable 'teacher' model — achieving near-teacher-level accuracy at a fraction of the size, latency, and cost.
Read definitionKokoro
An open-weight TTS model with just 82M parameters that matches the output quality of models ten times its size, running in under 0.3 seconds on CPU — the most efficient high-quality open-source voice model available.
Read definitionKV Cache Optimization
KV cache optimization is the set of techniques that shrink the memory footprint and bandwidth cost of the key-value cache Transformers rely on during autoregressive decoding. Left unmanaged, the KV cache grows linearly with sequence length, batch size, and model depth, and it is what actually limits how long a context window or how many concurrent requests a deployment can serve, not the model's parameter count.
Read definitionLarge Language Diffusion Models
A class of generative AI models that apply diffusion processes, originally popularized in image generation, to text generation. They offer non autoregressive, highly controllable generation by iteratively denoising continuous embeddings or discrete vocabulary spaces.
Read definitionLarge Language Model (LLM)
Large Language Models (LLMs) are advanced AI systems designed to understand, generate, and interact with human language at scale, leveraging vast amounts of text data to perform a wide range of linguistic tasks.
Read definitionStay ahead of the curve
Weekly newsletter on agentic AI, LLMs, and what we're building at Superteams — straight to your inbox.
Ready to ship AI in production?
We deploy fractional AI teams that deliver production-grade systems in 30–90 days. No fluff, no obligation.