Tech Feed - February 12, 2026

Feb 12, 2026

Articles and podcasts from the software engineering world.

Software Engineering Daily

Gas Town, Beads, and the Rise of Agentic Development with Steve Yegge

Duration: 1h 9m

Here's a comprehensive summary of the key points from the podcast episode:

Opening context:

The guest is Steve Yegge, a veteran software engineer exploring the frontier of "agentic software development" using AI tools like Beads and Gas Town.
The main topic is the evolution of AI-assisted programming, from chat-based assistants to full agent orchestration and multi-agent coordination.

Key discussion points and insights:

Yegge describes his own journey tracking the rapid progress of language models, from GPT-3.5 to the more recent Opus 4.5, and how each milestone represented a major leap forward in AI's coding capabilities.
He highlights the shift from "chat-oriented programming" (CHOP) to coordinating fleets of autonomous agents, which introduces new challenges around managing context, maintaining shared understanding, and orchestrating work across distributed systems.
Beads is presented as a task tracking system built on Git and SQL databases, providing a shared memory and coordination layer for AI agents. The transition to a Git-native database like Dolt is discussed as a key upcoming improvement.
With Gas Town, Yegge describes an experiment in multi-agent orchestration, where AI "polecats" and "crew" take on different roles and workflows, with the human maintaining more of a managerial/supervisory function.
A key challenge is keeping mental models of the codebase and the overall system up-to-date when the pace of development accelerates dramatically. Yegge discusses the need for technical leaders/advisors to help bridge this gap.

Notable technologies, tools, or concepts mentioned:

GPT, Opus, and other large language models
Beads (task tracking system built on Git and SQL)
Gas Town (multi-agent orchestration experiment)
Dolt (Git-native database)
Concepts like "chat-oriented programming" (CHOP), task graphs, shared memory, and "centaur" workflows blending human and AI capabilities.

Practical implications or recommendations discussed:

The need for developers to adapt their workflows and mentality as AI agents take on more of the coding work, shifting focus to managing context, orchestrating tasks, and maintaining shared understanding.
The emergence of new technical leadership roles focused on keeping the "big picture" of the system in mind, bridging gaps between human and AI capabilities.
Recommendations around prompting, bootstrapping, and managing AI agents, such as the importance of "landing the plane" with clear acceptance criteria.
The potential for tools like Beads and Gas Town to radically transform software development, but also the challenges in getting these systems to reliably operate at scale.

Overall, the discussion highlights the profound changes happening in software engineering as AI-driven development becomes more sophisticated and prevalent. Yegge provides a glimpse into the cutting edge of this transition, with valuable insights for developers, engineering leaders, and anyone interested in the future of software.

SE Radio

SE Radio 707: Subhajit Paul on ERP Automation and AI

Duration: 59 min | Read Transcript

Here is a comprehensive summary of the key points from the "SE Radio 707: Subhajit Paul on ERP Automation and AI" podcast episode:

Opening Context:

The guest is Subhajit Paul, who has over 20 years of experience leading ERP implementations at large global enterprises, particularly in electronics manufacturing and supply chain.
The main topic is how ERP systems work in the real world, including how they are implemented, where they tend to break, and how machine learning and generative AI are starting to be used within ERP.

Key Discussion Points and Insights:

ERP (Enterprise Resource Planning) systems integrate and automate core business functions like finance, procurement, production, inventory, and sales across an entire organization. This eliminates siloed processes and improves visibility.
The core ERP processes covered are order-to-cash, plan-to-produce, and procure-to-pay. These link customer orders to production, procurement, and invoicing.
ERP has evolved from on-premise to SaaS (software-as-a-service) delivery, and now is incorporating machine learning for predictive analytics and generative AI for more conversational, agentic capabilities.
Failures in ERP implementations can be chaotic, halting core business functions like production, procurement, and shipping. Proper planning, testing, training, and coordination are critical for successful rollouts.
Machine learning in ERP can enhance existing workflows, such as optimizing warehouse bin placement to improve picking efficiency. Generative AI is starting to add more flexible, conversational interfaces and autonomous decision-making.
Challenges exist in determining the right scope and granularity for AI agents, as well as when to use vendor-provided models versus building custom ones.

Notable Technologies, Tools, or Concepts Mentioned:

Machine learning models like k-means clustering, linear regression, XGBoost
ERP vendors and products: SAP, Oracle, Microsoft Dynamics
Emerging technologies: MCP (Model Context Protocol), A2A (Agent-to-Agent) communication
Vendor-provided AI capabilities: SAP Conversational AI (Joule), Oracle Fusion AI Agents, Microsoft Dynamics Copilot

Practical Implications and Recommendations:

Organizations need a clear data and AI strategy when implementing ERP, evaluating both vendor-provided and custom AI/ML capabilities.
The process for ERP implementations with AI integration should include steps for AI evaluation, data integration, model selection/training, and model integration.
Determining the right scope and granularity for AI agents within ERP processes is important, as is understanding when vendor models are sufficient versus needing customization.
Integrating AI/ML into ERP can enhance existing workflows, but organizations must also be prepared to extend or modify vendor-provided AI capabilities to meet their specific business requirements.

Overall, this episode provides a comprehensive overview of how ERP systems work, the challenges and evolution of ERP implementations, and the growing role of AI and machine learning within ERP platforms. It offers practical insights for organizations looking to leverage these technologies to improve their core business processes.

The Cloudcast

Three AI Rooms to be a Fly on the Wall in 2026

Tech Brew Ride Home

How To AI With WSJ's Chris Mims

Duration: 52 min | Read Transcript

Here is a comprehensive summary of the podcast episode "How To AI With WSJ's Chris Mims":

Opening context:

The guest is Chris Mims, a technology columnist for the Wall Street Journal.
The main topic is Mims' new book "How to AI: Cut Through the Hype, Master the Basics, Transform Your Work", which aims to provide a practical guide for non-technical people on how to effectively incorporate AI into their workflows.

Key discussion points and insights:

Mims prefers the term "simulated intelligence" over "artificial intelligence" to highlight that current AI systems, while capable, lack the full breadth of human intelligence.
He describes AI as akin to "a super intelligent toddler" - capable of remarkable feats, but still needing guidance and oversight.
Mims shares his own journey from AI skeptic to enthusiastic adopter, highlighting tools like Notebook LM that dramatically improved his research and writing workflows.
He emphasizes that AI tends to benefit experts the most, as they can better evaluate the AI's outputs and ask the right prompting questions.
Mims discusses examples of AI being used effectively in fields like law, construction, and consumer packaged goods - automating tedious tasks and augmenting human expertise.

Notable technologies, tools, or concepts mentioned:

Notebook LM, Copilot for Depositions, and other AI-powered research and writing assistants
The concept of "machine psychology" - understanding AI's behavior and quirks rather than just the underlying math
The "laws of AI" outlined in Mims' book, such as "experts benefit the most from AI" and "give it your least favorite tasks"

Practical implications and recommendations:

Start small by having AI handle your least favorite, most tedious tasks first to see the benefits
Embrace an experimental mindset and be willing to try new AI tools, rather than sticking to old workflows
Leverage AI's strengths to enhance your productivity, but be wary of cognitive offloading and losing essential domain expertise
Consider using AI to do "less" work by automating annoying tasks, rather than always striving for "10x" productivity

Overall, the episode provides a nuanced, experience-based perspective on the current state of AI and how non-technical professionals can effectively harness it to transform their work, while avoiding the pitfalls of hype and over-automation.

Anthropic Makes Its Move

Duration: 19 min

Anthropic raises the second largest financing round of all time. Other AI players are beginning to show hockey stick revenue growth. Meta wants to add facial recognition to its glasses. Ring pulls back from some recognition partnerships for its camera. And, of course, your Weekend Longreads Suggesti...

Pour Moi, C'est Le Déluge

Duration: 21 min

Aaannddd…. Right on time here come the Chinese AI models. Elon Musk kicks off a major reorg of xAI. Google is warning of AI distillation attacks. New Waymo cars hit the road. And another interesting AI essay to read to you. Chinese AI startup Zhipu releases new flagship model GLM-5 (Reuters) Musk ...

The “Covid Moment” For AI?

Duration: 21 min

Forget the ChatGPT moment, is this the “Covid Moment” for AI, the moment when everything has changed its just that not everyone knows it yet? More weird musical chairs at the AI companies. TikTok but for around the block. And do we finally have a universal translator but for phone calls? ⁠Something ...

Dev Interrupted

Breaking GitHub, AI vampires & the great Oz | Warp’s Zach Lloyd

Duration: 30 min

Did AI agents just DDoS GitHub? Andrew and Ben are joined by Warp Founder and CEO Zach Lloyd to discuss the massive strain agentic workflows are putting on our infrastructure and why the "Monday Morning Commit Spike" is the new normal. They also dive into Steve Yegge’s reflective piece on ...