2026-09 - W4

Built with Jev | DAIR.AI Academy | DAIR.AI Academy

Interesting examples of using jev.

I am often wrong - Boris Cherny’s Blog

Sometimes I will give feedback to people when they are missing steps in the framework, or are poorly executing some of the steps. I expect the same feedback in return. I try hard to give the feedback in real time, so the person/team can learn more quickly. Most often, the failure mode I see is (3) failure to clearly define the problem, and (4) failure to define an approach that is clear and simple. When one of these is missing, it leads to complex plans and unclear success criteria. For complicated problems, lack of clarity can be hard to spot if you’re the one making the plan, making it even more important to get feedback from people.

I Don’t Want to Read What You Didn’t Write

I don’t want to live in a world where you use AI to summarize something important into unreadable text, and then I use AI in an attempt to decipher it. I want to hear you, imperfections and all. I want your interpretation of aesthetics, beauty, quality, relationship, time. I want to know how you feel. I want you to cut through and tell me what really matters. I don’t want to read a summary of facts and connections separate from experience and relationship. … Intentional writing will likely become more valuable. People who write, and write to think, to think deeply and carefully, or to create, to share, or to capture something important without explicitly expressing it will continue to write and produce original work. The people who never were writers will use AI to produce lots of text.

Don’t use AI to write | Paul Bakker

Writing requires thinking, don’t let an AI do the thinking for you. Writing forces you to deeply think about the problem you’re solving, and what you’re trying to communicate to your readers. This is why writing is hard, but also why writing is such a useful tool to help you think. If you generate a document based on some bullet points, or whatever the source is, you’re skipping the thinking step. The result probably looks “pretty good”, and you’re likely going to accept it with some minor tweaks. But did you think deeply about the problem? Does the document really reflect the most important things you wanted to communicate? Probably not. But you’ll never know, because you skipped the thinking step.

We write code by hand: We use Natural Intelligence to build software.

Jev in 25 lines of Python - NobodyWho

Clankers Made Me Build a Second Brain | Jadarma’s Blog

To inoculate yourself don’t delegate understanding. If you build your own understanding you will be the one who earns the dividends.

I am sorry, but everyone is getting syntax highlighting wrong @ tonsky.me

Yes, aligned with his reasoning. That’s why I have use nvim-grey. Some of his ideas are worth exploring:

Comments should be highlighted, not hidden away. Use background colors!

Healthy Feedback

• Obnoxious aggression. Challenging directly without showing personal care.
• Ruinous empathy. Caring without challenging directly.
• Manipulative insincerity. Neither caring nor challenging directly.

How to Unclench | Jonny Miller

Interesting read.

First Principles Thinking · Sunil Sadasivan

First principles thinking makes working that way feel natural. When you truly understand what you’re trying to accomplish, it’s easier to take a small step, learn from it, and keep going. If you’re doing it right, working with agents lets that back-and-forth happen much faster. You get faster learning loops and more momentum, oriented around deep understanding. To me, that’s the new flow state.

Long Live Human Thinking.

Ollaya · Run decision models locally

jev locally.

GitHub - firelex/jeff: Fine-tunes of Qwen3.5 and Gemma 4 for zero-shot classification · GitHub

Another one that also seems promising.

AI Agents Push Humans Out of the Loop

A commonly proposed solution is human oversight and keeping a ‘human in the loop’, but this is not a simple solution: Not only do current approaches to AI agent design impede effective human oversight, but the cognitive capacities required for it are also themselves degraded by extended use of AI systems.

AX

Declare an agentic task. AX runs it at scale.

AX sandboxes your task, wires up its workspace, fences its network, and helps you run a large number of them per cluster. Either use a single task per agent, or compose as many as your agent needs.

Planning with Agents: Divided Worlds, Boundary Objects, and Thicker Interfaces

Really good essay (even if at the end, it was to promote her work for Github Next). She highlighted the interface issue when doing plan, where wall of texts are not the right medium. This made me think of plannotator to improve the harness experience with human.

10 tells of a slop ui

Postgres AT TIME ZONE ‘UTC’ does NOT do what you think it does

Now you are inadvertently using the timestamp without time zone (aka timestamp) data type, which is markedly discouraged.

A staff engineer’s guide to inventing work

I Built a Local AI Music Studio - DEV Community

Might be useful for me to generate music from the kids’ poems, so they can learn poems more easily.