text stringlengths 0 2.68k |
|---|
**Bert Hubert:** Yeah, so that's an interesting one. So I've been pretty enjoying my time on Mastodon. I think it's really a geek heaven. So that works really well. But interestingly, people ask me that question quite a lot, and I've found, and I think it's a very weird, strange sort of regression... Many people are no... |
And I've found it to be a bit pompous, to write mail updates on what Bert has been doing... But apparently, there is demand for that. So I am probably gonna restart a small newsletter. |
**Jerod Santo:** Nice. |
**Bert Hubert:** That kind of stuff. I also used to tell people to just follow me on LinkedIn, but the problem there is that, on one hand, I'm sort of active in Dutch politics, things which are entirely not interesting to the rest of the world... |
**Jerod Santo:** To a lot of people, sure. |
**Bert Hubert:** Yeah. And the other way around, there are many people there that follow me for the Dutch politics stuff, and they are sort of confused, they are massively confused to suddenly see me talk about -- |
**Jerod Santo:** \[laughs\] "What's all this nerdy stuff?" |
**Bert Hubert:** C++... And then they go "What is this?" So actually, for the newsletter thing I'm building now, it's actually gonna have like four newsletters. It's gonna be one for people that like Dutch news, that like news in English, that like stuff about politics, and that like stuff about technology. And you can... |
**Jerod Santo:** Nice. Gotta love RSS. Here's a name you can name your new newsletter... You can call it Bert's Pompous Update. \[unintelligible 01:25:30.16\] |
**Bert Hubert:** Yeah. "Bert's self-important things that you actually all asked for." Yeah, just tongue-in-cheek that sucker. It always works. Cool. Love it. We appreciate you coming on the show. Nice meeting you. Adam, anything else you wanted to talk about or ask? |
**Adam Stacoviak:** Keep that software simple. |
**Bert Hubert:** Yeah, that's the goal. That's the goal. |
• Discussion about Zed's position in the market amidst new editor tools |
• Nathan Sobo's perspective on being GPL-ed, making it difficult for others to fork Zed |
• The challenges and opportunities of building AI-integrated software, including potential competition from forking VS Code |
• Adam Stacoviak's humorous commentary on the golden retriever on acid analogy |
• Nathan Sobo's plans for future development, including a debugger and AI integration |
• The tension between being competitive in the market and maintaining independence as an open-source project |
• Development of a product that integrates a stochastic, unpredictable process |
• Challenges of testing and debugging a system with inherent unpredictability |
• Approaches to testing a system with flaky, non-deterministic behavior |
• Integration of a new technology with existing software engineering practices |
• Challenges of working with a Large Language Model (LLM) and its unpredictable output |
• Need for a more tolerant parser to handle the LLM's output |
• Comparison of working with the LLM to working with network latency and unpredictability |
• Discussion of the prompting mechanism and its unique challenges |
• The challenges of working with large language models (LLMs) and the difficulty of testing and debugging their behavior |
• The importance of open-source development and transparency in software development |
• The use of simulated random network latencies and custom schedulers in Rust to test and debug LLMs |
• The problem of legacy codebases and the need for well-tuned, high-quality software systems |
• The potential of AI to simplify software development and make it more efficient and effective |
• The need for rigorous design, testing, and maintenance of software systems, and the benefits of using AI to support this process |
• The ease of creating and automating mundane tasks with AI-powered tools. |
• Concerns about the potential loss of "purity and greatness" in software development due to increased speed and efficiency. |
• The idea that the world may not respect traditional software development standards and values. |
• The potential for AI to replace human developers and create a world where software development is like a hobby. |
• The increased importance of testing and quality control in a world where automation is making it easier to create software. |
• The need for developers to raise the bar on software quality in the face of new technology. |
• The changing value proposition for software development, with a greater emphasis on sustainability and progress rather than just speed. |
• The founders of Zed and Cursor discuss the development timelines and strategies of their respective projects |
• Nathan Sobo of Zed feels they are not late to market, but rather part of an early wave of AI-focused code editors |
• Zed's focus on core editor functionality allowed them to lay groundwork for future integrations with AI and LLMs |
• OpenAI is reportedly in talks to acquire Windsurf (formerly Codeium) for $3 billion |
• Nathan Sobo believes the value of their technology lies in its ability to integrate LLMs with end-user experiences |
• The future of software development may involve a shift towards more collaborative, agent-driven workflows |
• The role of traditional tools like Git in a world with AI-driven code editors and workflows is uncertain |
• Augmenting Git with a finer-grained understanding of changes |
• Editor as a vertically-integrated authoring experience |
• Agentic coding experience with LLMs in Zed, including chat and code editing |
• Multi-buffer experience for viewing and editing multiple files at once |
• Reusing multi-buffer concept in Git and search results |
• Improving code review and diagnostics with multi-buffer features |
• The concept of a multi-buffer, where multiple files are edited simultaneously in a single virtual buffer |
• The idea of using AI agents to propose edits and the importance of disentangling human and AI edits |
• The potential for persisting and recalling edited buffers, especially in the context of AI-assisted collaboration |
• The technical challenge of creating persistent and fine-grained data structures to track edit history |
• The user experience of interacting with the AI agent, including using keyboard shortcuts to access the agent |
• The potential for AI-assisted collaboration to revolutionize the way developers work |
• Improving YAML files using an LLM |
• Collaborative editing with an LLM |
• Stashing and unstashing code |
• Signed commits and authentication |
• Experience with Sonnet 3.7 and its improvements |
• User experience and making tools magical |
• Comparing Zed to VS Code and "flying coach" |
• The concept of taking for granted the conveniences of modern technology, such as Wi-Fi on a plane, was discussed in relation to a Louis CK quote. |
• Nathan Sobo discussed how he tends to complain about things he has the most responsibility and ability to impact, and how he tries to phrase his complaints in a positive way. |
• Jerod Santo mentioned how he appreciates the weekly updates and features added to a product, even though he didn't contribute financially. |
• The conversation turned to discussing the user experience of a product, with Adam Stacoviak praising the clarity and polish of the interface. |
• Nathan Sobo acknowledged the team's hard work in creating a good experience and expressed gratitude for the feedback. |
• The key goal is to create an experience that feels immediate and engaged, similar to watching a stream. |
• Nathan Sobo discusses the benefits of building on top of a well-established code editor, such as VS Code, for developers. |
• He shares his approach to building Zed, focusing on creating a tool for software engineers, not just developers, with an emphasis on integrity and speed. |
• Sobo mentions the importance of understanding how the software being relied on works, and how LLMs can provide leverage for software engineers. |
• The conversation turns to the future of Zed, with the goal of making it accessible and approachable for new users, while also addressing the needs of experienced software engineers. |
• The team discusses the potential for users to create new projects with Zed, such as building a Rails app, and the agents' ability to execute command-line tools. |
• Nathan Sobo mentions that the agents run in the user's shell, allowing for a seamless integration with the user's environment. |
• Integrating model context servers with Zed for more customized workflows |
• Adding support for image-based interactions, such as taking screenshots |
• Development of tools for tasks like Postgres schema manipulation and web browser automation |
• Emphasis on extensibility and customizability in Zed's interface |
• Discussion of the benefits of "vibe-coding" or "collab-coding" in developing complex workflows |
• Sharing of experiences with using Zed for tasks like editing shader code and exploring settings files |
• Discussion of a software feature request for light and dark modes |
• Adam Stacoviak's solution to dark mode issue using a JSON file |
• Launch of Zed AI, a platform with AI features such as write-optimized editing and inline assist |
• Edit Prediction feature and its integration with Subtle Mode |
• Discussion of multiplexing algorithmic and model-based suggestions |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.