RT @jrdothoughts: I really enjoyed writing this one :) The Sequence Engineering #676: Hacking with Gemini CLI

RT @layerlens_ai: You asked us to take a closer look at @xai’s Grok 3 – and here’s what we found 👀. It dominates enterprise tasks like data….

RT @jrdothoughts: Some of my initial observations about @eigenlayer Eigen Cloud: The Verifiable Cloud: Some Notes About EigenCloud https://….

RT @layerlens_ai: 🚨 New blog drop: In Focus: DarkBench | Issue 3.We dig into behavioral safety evals and why benchmarks like DarkBench are….

RT @layerlens_ai: We’ve just launched our official YouTube channel 🎥. Catch product demos, walkthroughs, and practical tips for getting the….

RT @layerlens_ai: We have onboarded @MistralAI's Magistral Medium Models: both the normal and the thinking variant. Early results put it….

This is a week old, but figured I will share here. Highlighting some safety work I had the pleasure of being a part of recently. We used steering vectors to influence the behavior of open source models, making their responses more honest. Read more here:

RT @jrdothoughts: Some thoughts about the myths and realities of web3-AI and a big of a wake up call. @PluralisHQ , @NousResearch , @PrimeI….

RT @layerlens_ai: Small shoutout in @CoinDesk today 👀. LayerLens was briefly mentioned among teams working on foundational challenges in We….

Going to be in SF this week. Who should I meet?. Reply @here.

RT @layerlens_ai:

RT @layerlens_ai: Our independent analysis on @NVIDIAAI's #Nemotron models is now live. Use our dashboard to get full comparisons of how….

RT @layerlens_ai: How well does DeepSeek-R1-0528 actually understand accounting?. We put it to the test on our Accounting Audit benchmark.….

RT @layerlens_ai:

RT @layerlens_ai: 🚨 Don’t miss it!.Join us for “The Good, the Bad & the Surprising: Lessons from AI Evaluation” – our first ever webinar o….

RT @layerlens_ai: 🚀 Big news: @layerlens_ai is partnering with @Conformiq_inc to bring robust, real-world benchmarking to enterprise AI tes….

RT @layerlens_ai: @GoogleAI released a new checkpoint for @GeminiApp 2.5 Flash at #GoogleIO yesterday. Some results from the LayerLens Atl….

RT @layerlens_ai: What really happens when you put today’s top AI models to the test?. Join us for a live webinar where we reveal the most….

Our goal is to transform how we think about benchmarks and evals for models, agents, and more. A great article by my co-founder and president @jrdothoughts on our vision:.

Consistent, independent benchmarks are probably one of the most poignant problems in the space. Atlas not only lets you view benchmarking results, but also gives you full transparency, along with an entire suite of analytics. Try it out here: