Auto Added by WPeMatico
Perplexity has open sourced Lily, the local inference engine behind Hybrid Compute in Perplexity Computer. It is a single-process runtime: a Rust layer loads the checkpoint and drives the generation loop, an OpenAI-compatible chat-completions API streams tokens, and hand-written Metal kernels execute the model. Neither PyTorch nor MLX sits in the execution path. Lily is
Coding agents spend a large share of their tool budget on search. When the target is a known symbol, ripgrep answers it exactly. When the target is a behavior described in plain language, keyword matching often misses, and the agent falls back to guessing terms, reading whole files, and assembling context by hand. Each of
Google just announced Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, three weeks after Gemini 3.7 Flash and marking the third Flash release in six weeks. Both variants run on the same foundational intelligence, refined through long-running agentic loops that recursively evaluate the underlying models. What separates them is not architecture, it is the safety
Enterprise AI buyers have been stuck between two things they both need. Regulated teams need a zero data retention (ZDR) guarantee, so no prompt or agent transcript sits on a vendor’s servers. Security teams need misuse detection, which historically required the vendor to hold that same data long enough to correlate it. This week, Anthropic
Most production voice stacks are three systems stitched together. One model transcribes, a second separates speakers, and a detector decides when the user stopped talking. Each hand-off adds latency and a new failure mode. Muse Voice Transcribe, announced by Meta Superintelligence Labs this week, collapses those three jobs into a single autoregressive model. Meta calls
Agentic assistants have a structural problem: the context that makes them useful — deal documents, privileged files, client records — is exactly the context users cannot send to a cloud endpoint. This week, Perplexity shipped its answer for Mac. Hybrid compute splits a single Perplexity Computer task between frontier models in the cloud and a
Anthropic has released Claude Fable 5.1 and Claude Mythos 5.1, three months after the Fable 5 line shipped in June 2026. The two are the same underlying model behind different safeguard layers. Fable 5.1 is generally available as claude-fable-5-1; Mythos 5.1 stays restricted to vetted organizations. Both carry a 1M token context window and 128K
Quantitative research agents that write their own experiments can corrupt the evidence they later learn from. A leaky feature that scores well gets stored as a successful precedent and propagated through later iterations. Prompt-level instructions and reviewer agents do not close this, because author and reviewer share the same blind spots. A team of researchers
“It’s not a technology problem. It’s a process problem.” “It’s not an accuracy issue. It’s an adoption issue.” “It’s not the model. It’s the human.” Sound familiar? I hear these arguments all the time. But here’s the reality: generative AI has an accuracy problem. In fact, it is well documented […] The post Stop saying
Artificial intelligence (AI) can review vast financial datasets, identify unusual transactions, and extend control testing across entire populations. Yet an alert does not explain intent, business context, or regulatory significance. An atypical journal entry may indicate misconduct while also reflecting a legitimate exception the model has never encountered. That distinction reveals the central limitation of
Why Human Judgment Is Essential in AI-Powered Financial Controls Read More »
