Machine learning & AI podcaster, community builder and all around enthusiast. Creator of the @TWIMLAI Podcast, TWIMLcon, TWIMLfest & the TWIML Solutions Guide.
St. Louis, MO
Joined April 2007
- Tweets19.2K
- Following2.8K
- Followers18.9K
- Likes15.3K
Sam Charrington retweeted
The best interview I’ve listened to all week, by @samcharrington with @CompleteSkeptic twimlai.com/podcast/twimlai/…
A few years ago we were talking about a severe shortage of data scientists. Now @googlecloud is casually demoing a Gemini Agent building an ML model and training it on Spark #GeminiAtWork
Watching the #GeminiAtWork event for the latest updates in enterprise agents from @googlecloud
Key news:
- Gemini Agent is new agentic experience in Google Workspace.
- Agents now have persistent identities that users can interact with in Chat, Docs, comments, desktop, web, devices, etc.
- Admins govern user & agent interactions in console, and apply policies and budget across the org.
Sam Charrington retweeted
@samcharrington asks @CompleteSkeptic questions about Jev you won't hear from other (fawning) podcast hosts. Super informative. Slightly less convinced 'its just a classifier' but remain skeptical on economics/adoption.
In this episode, @CompleteSkeptic, co-founder and CEO of @typesafeai, joins us to discuss Jev, TypeSafe’s recently released model for bringing fast, reliable intelligence directly into software. We explore the idea of “machine-native intelligence” and why Diogo believes models optimized for generating text are poorly suited to many of the decisions required for real-world automation. He explains how Jev differs from traditional classifiers and LLM-based approaches, the role of reinforcement learning from calibrated decisions (RLCD), and why calibration and reliability are central to making AI useful as a software primitive. We also discuss the relationship between models and code, why Diogo believes AI systems should become more engineered rather than relying on a single model to do everything, and how Jev-like models could reshape agents, tool use, and the architecture of AI-powered software.
🗒️ Full show notes: twimlai.com/go/779.
📖 CHAPTERS
===============================
00:00 - Introduction
01:58 - Hot Takes
07:42 - The Bitterest Lesson
12:25 - Diogo’s Career Background
13:58 - Sci-Fi Software Demos vs. Easy Unsolved Problems
21:13 - Jev
22:37 - How Jev Differs from Classifiers and Small Models
32:20 - Intelligence and RLCD
35:57 - Jev as an Intelligence API
38:47 - RLCD Post-training
46:01 - RLCD and Calibrated Decisions
54:06 - Uncertainty Estimation
57:43 - Generalization
01:08:25 - Models, Code, and Harness
01:16:28 - Jev and the Future of AI Agents
01:22:21 - Rigid Code Workflows vs. Dynamic Agent Loops
01:27:05 - What’s Next
At #SnapdragonSummit Vinesh Sukumar shares double digit generation-to-generation increases in @MLPerf benchmarks in both performance and efficiency @MLCommons @Qualcomm
Sam Charrington retweeted
Agree on these points @RolandMemisevic @NaveenGRao . I have a convo with @GregHBurnham coming to the @twimlai pod next week where we dig into this.
.@clattner_llvm takes the #SnapdragonSummit stage for the first time as EVP Advanced AI Software @Qualcomm following the @Modular acquisition.
Introducing the audience to the stack and it's role in the broader ecosystem.
First up day 2 at #SnapdragonSummit, a fireside chat with @cristianoamon CEO @Qualcomm and @ramin_m_h ceo @liquidai talking about co-designing hardware, models, harnesses and agents for efficient on-device AI at scale.
.@ramin_m_h announcing Liquid Context on Snapdragon a memory layer for on-device agents with read, compress, write capabilities.
• With your permission, device signals become an understanding of your routines, preferences and needs.
• That context helps your chosen agents: local, cloud or across both.
• Liquid Agent: our efficient embedded agent, powered by LFM2.5-2.6B and optimized for Hexagon.
• OEMs get a shared foundation for personal context, plus a strong embedded agent they can choose to build with.
@Qualcomm #SnapdragonSummit
Mobile inference capabilities of @Qualcomm #Snapdragon processor and Hexagon NPU continue to grow. Now supporting 30B+ parameter MoE models @cpatrickSD #SnapdragonSummit
Sam Charrington retweeted
.@cpatrickSD says that @Snapdragon is the processor for the agentic age. #SnapdragonSummit #Paditravel_Snapdragon
.@Qualcomm CEO @cristianoamon laying out his vision for the AI Smartphone, the centerpiece of AI and agentic experiences. The device is key bc it can understand your goals and intent, and has your context.
#SnapdragonSummit
.@Qualcomm scaling its High Bandwidth Compute memory architecture, designed for the data center, down to notebooks, smartphones, smart glasses powered by #Snapdragon @cristianoamon
Excited to be here @Qualcomm #SnapdragonSummit kicking off now with CMO @donnymac. AI agents to be a recurring theme at this year's event.
In the past week according to my X feed we’ve cured Alzheimer’s three times, having separately determined that it is due to microplastics in the brain, mouth bacteria in the brain, and a mineral deficiency.
I’m hopeful that as AI accelerates science and medicine we’ll see real breakthroughs at an unprecedented pace. But we’re going to need much better ways of distinguishing those breakthroughs from the daily stream of papers being promoted as breakthroughs.
This quoted post is unavailable.
Sam Charrington retweeted
How me and Claude work together
xitter.cf/ravikiran_dev7/status/…
Sam Charrington retweeted
can LLMs replace embedding models? in a new work "embedder's dilemma" we find LLMs now beat embedding models -- but at much higher cost. when to choose which?...
📜 arxiv.org/abs/2608.12875
