Subvolts take OpenAI's new GPT models and Mistral's open-source release highlight the fast-paced AI race. Mar 26, 2026 - Developers
- Learn about new model releases like GPT-5.4 and Mistral Small 4, plus agent runtime tools from Nvidia.
- Business
- Understand OpenAI's enterprise pivot, Mistral's Forge for custom models, and the competitive agent OS landscape.
This episode covers significant product announcements from OpenAI and Mistral, and discusses industry shifts in agent technology and enterprise AI.
#238 - GPT 5.4 mini, OpenAI Pivot, Mamba 3, Attention Residuals
Audio streams from Skynet Today. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Andrey Kurenkov and Jeremie Harris discuss AI news on Last Week in AI. OpenAI launched GPT-5.4 mini and nano with large context windows. Mistral open-sourced its Small 4 model family and introduced Forge for custom model training. The episode covers intensified agent OS competition, business strategy shifts at OpenAI and Microsoft, and new AI safety research.
AI labs are releasing new tools. OpenAI has new versions of its AI with much larger memory, while Mistral is sharing its AI models for others to use. This helps businesses build their own AI and shows how companies are competing to create smarter AI assistants.
This episode offers a comprehensive overview of recent AI developments, crucial for understanding the competitive landscape. It's for AI professionals, business leaders, and enthusiasts tracking frontier lab strategies. Listeners will learn about OpenAI's product expansion, Mistral's open-source contributions, and the evolving agent ecosystem. The discussion highlights how major players are adapting to market pressures and advancing AI capabilities, providing context for ongoing industry shifts and safety considerations.
- HostAndrey Kurenkov
- HostJeremie Harris
- Episode Title: #238 - GPT 5.4 mini, OpenAI Pivot, Mamba 3, Attention Residuals
- Show: Last Week in AI
- Length: 121 minutes
- OpenAI Models: GPT-5.4 mini and nano with 400k-token context windows
- Mistral Models: Small 4 family (MoE, 119B total/6B active)
- Agent OS Competition: Meta's Manus, Nvidia NeMo/Open Shell
Our 238th episode with a summary and discussion of last week's big AI news!
Recorded on 03/18/2026
Hosted by Andrey Kurenkov and Jeremie Harris
Feel free to email us your questions and feedback at andreyvkurenkov@gmail.com and/or hello@gladstone.ai
Read out our text newsletter and comment on the podcast at https://lastweekin.ai/
In this episode:
* OpenAI released GPT-5.4 mini and nano with 400k-token context windows, higher per-token prices but claimed token-efficiency gains in Codex; nano is API-only and pitched for high-volume classification/data extraction despite a major price increase.
* Mistral open-sourced the Small 4 model family (MoE, 119B total/6B active) combining reasoning, multimodal, and coding-agent capabilities, and announced Forge to help businesses train or post-train custom models.
* Agent “operating system” competition intensified with Meta’s acquired Manus launching a local Mac agent, Nvidia announcing NeMo/“Open Shell” sandboxed agent runtime, and Nvidia also unveiling DLSS 5 plus major hardware forecasts including Groq LPU integration.
* Business and safety updates included OpenAI shifting focus toward productivity/enterprise amid competition, Microsoft reorganizing Copilot and frontier-model efforts, Meta delaying its next model, China-linked ByteDance deploying large Nvidia clusters abroad, and new safety work on steganography, chain-of-thought faithfulness, fine-tuning defenses, cyber-attack evals, and constitution/spec compliance.
A thank you to our current sponsors:
- Box - visit Box.com/AI to learn more
- ODSC AI - go to odsc.ai/east and use promo code LWAI for an additional 15% off your pass to ODSC AI East 2026.
- Factor - head to factormeals.com/lwai50off and use code lwai50off to get 50 percent off and free breakfast for a year
Timestamps:
- (00:00:10) Intro / Banter
- (00:01:56) News Preview
- Tools & Apps
- (00:02:39) OpenAI ships GPT-5.4 mini and nano, faster and more capable but up to 4x pricier
- (00:08:04) Mistral's new Small 4 model punches above its weight with 128 expert modules
- (00:14:03) Meta's Manus launches 'My Computer' to turn your Mac into an AI agent - 9to5Mac
- (00:17:57) NVIDIA Announces NemoClaw for the OpenClaw Community | NVIDIA Newsroom + Nvidia boosts knowledge work with Open Agent Development Platform
- (00:24:09) DLSS 5 looks like a real-time generative AI filter for video games | The Verge
- (00:26:36) OpenAI to Launch Chat
- context windows
- Andrey Kurenkov, Jeremie Harris
- how much information an AI can consider at once
- token-efficiency
- Andrey Kurenkov, Jeremie Harris
- how well an AI uses its processing power per piece of information
- MoE
- Andrey Kurenkov, Jeremie Harris
- Mixture of Experts, a way to make AI models more efficient
1 h 23
1 h 15
1 h 011 h 03Questions people ask
- What are the new OpenAI GPT models?
- OpenAI released GPT-5.4 mini and nano, featuring 400k-token context windows and claimed token-efficiency gains.
- What did Mistral announce?
- Mistral open-sourced its Small 4 model family and launched Forge to help businesses train custom AI models.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Skynet Today.