Transformers

Why Agent Hype can fall short of reality – Joel Becker, METR

Dec 24, 2025 · 21:22

Joel Becker, a researcher at METR, examines why AI models that ace benchmarks like SWE-bench fail to boost real-world developer productivity. METR's time horizon measurements show AI capabilities doubling every 6-7 months, with Claude 3.7 Sonnet achieving a 50% success rate on tasks taking humans 4 minutes. Yet a randomized controlled trial of 16 experienced developers on large open-source projects found they were 19% slower when using AI tools like Cursor Pro, contradicting expert predictions of 40% time savings. Becker attributes the gap to high context requirements, low AI reliability, and task complexity—developers spent significant time verifying and correcting AI outputs. The episode warns that benchmark-style evidence, which uses low-context human baselines, overstates AI readiness for messy, interdependent real-world tasks.

Unveiling the latest Gemma model advancements: Kathleen Kenealy

Feb 9, 2025 · 16:25

Kathleen Kenealy, technical lead of the Gemma team at Google DeepMind, unveils the latest advances in the Gemma model family, including the launch of Gemma 2 in 9B and 27B parameter sizes, which outperform models two to three times larger, such as LLaMA 3 70B. She also introduces PALI Gemma, a multimodal model combining Siglip Vision Encoder with Gemma 1.0 for image-text tasks. The episode highlights Gemma's responsible-by-design approach, broad framework support (TensorFlow, Jax, PyTorch, etc.), and the release of the Gemma cookbook with 20 recipes. Kenealy emphasizes that Gemma 2 is optimized for easy integration and fine-tuning, available on Google AI Studio, and invites the community to build and share their projects.

LLM Safeguards: Security Privacy Compliance Anti Hallucination: Daniel Whitenack

Dec 31, 2024 · 34:10

Daniel Whitenack of Prediction Guard outlines a practical checklist for deploying secure and accurate LLMs in enterprise, addressing hallucination, supply chain vulnerabilities, flaky model servers, data breaches, and prompt injection. He proposes a factual consistency model fine-tuned to detect inconsistencies between AI output and ground truth, rather than relying solely on RAG. For supply chain risks, he recommends a trusted model registry and industry-standard libraries. Data breaches are mitigated via PII detection filters and confidential computing (e.g., Intel SGX) to encrypt server memory. Prompt injection is countered with a custom firewall layer using ensembled classification models. In Q&A, he discusses latency trade-offs using smaller NLP models, pre-production testing for visibility, data access controls via role-based database queries, SIEM integration for monitoring new artifacts like model caches, and additional challenges with agents such as excessive agency, proposing dry-run approvals.

No more bad outputs with structured generation: Remi Louf

Oct 14, 2024 · 15:32

Rémi Louf, CEO of .txt and co-maintainer of Outlines, argues that structured generation—guiding LLMs to produce valid regex, JSON, or context-free grammar outputs—eliminates parsing errors and hallucinations while adding negligible overhead. Outlines masks tokens that violate the target structure, enabling Mistral 7B to achieve 99.9% valid JSON (vs. 17% without). It also accelerates inference: a chatty 50-token ChatGPT answer shrinks to 8 tokens, and structured generation boosts open models beyond GPT-4—Phi-3 Medium hits 96.5% on Berkeley function calling (GPT-4 gets 93.5%). With one shot matching eight shots in accuracy, Louf contends that most text is structured and that structured generation should be the default for non-chatbot workflows.

Everything you need to know about Fine-tuning and Merging LLMs: Maxime Labonne

Sep 25, 2024 · 17:52

Maxime Labonne from Liquid AI explains the LLM training lifecycle—pre-training, supervised fine-tuning (SFT), and preference alignment—and when to use fine-tuning over prompt engineering. He details SFT dataset creation (accuracy, diversity, complexity) and techniques like full fine-tuning, LoRA, and QLoRA, with key hyperparameters. The core of the talk is model merging: combining weights of fine-tuned models without GPU, using methods like SLERP (spherical linear interpolation for two models), TIES (pruning redundant parameters to merge many models), pass-through (concatenating layers, e.g., Meta LLaMA 3 120B Instruct by repeating layers gives strong creative writing), and Franken-MoE (extracting FFN layers from domain-specific models with a router). Labonne demonstrates these with his NeuralBeagle and Beyonder models, noting merged models dominate the OpenLLM leaderboard and that TIES merging often outperforms more experimental Mixture of Experts approaches.