Eleanor Berger
banner
intellectronica.net
Eleanor Berger
@intellectronica.net
🌐 https://intellectronica.net/

👩‍💻 Expert AI Leadership (ex google microsoft startups ...)

👩‍💼 Consulting https://okigu.com/

👩‍🏫 Teaching http://nanolink.xyz/ai-coding
Honest but challenging question about DSPy: if it's really so good, how come it's not part of many consumer products?

End-users of AI want nothing more than getting better results without having to learn how to prompt. If DSPy is the solutions, why isn't it heavily productised?
November 11, 2025 at 1:14 PM
I know I'm late to the party, but I only recently discovered this so sharing in case anyone else still needs an update.

You can update more-or-less everything from NodeJS to Bun and it works the same, just _much_ faster. It's a majoy QoL improvement.
November 11, 2025 at 9:07 AM
November 11, 2025 at 7:29 AM
Claude Code rocks for fast implementation, but GPT-5-high crushes analysis & planning! 🚀 Command: /codex to prompt it from Claude, get detailed plans, then build with Sonnet/Haiku. elite-ai-assisted-coding.dev/p/gpt-5-hig...
Analyse and Plan with GPT-5-high from Claude Code
Claude Code with Claude Sonnet 4.5 and even Claude Haiku is fast, smart, and fun to work with.
elite-ai-assisted-coding.dev
November 10, 2025 at 6:21 PM
Just read that Sarkozy was released home and I'm happy about that, even though he's a douchebag. Putting people in prison is stupid and vengeful. It should be reserved to cases where this is strictly necessary for protecting others. Punishment is medieval.
November 10, 2025 at 3:24 PM
⭐⭐⭐⭐⭐ feedback from the first cohort of Elite AI-assisted Coding.

@isaac_flath and I built a course that brings together the frameworks and techniques for applying AI in every aspect of software development. It worked. 👌

👉 Join us for cohort 2: nanolink.xyz/ai-coding-c...
November 10, 2025 at 2:54 PM
Early stopping is a brilliant optimisation and a huge risk. In human learning (not just ML).

Are you done, or just stuck at a local maximum?

The world is changing fast. If you don't regularly challenge what you think is optimal, you may be at the tallest point in the valley.
November 10, 2025 at 1:10 PM
Managing my co's expenses is never fun, and there are always some complications (like two invoices in one scanned PDF).

@warpdotdev with Skillz ( github.com/intellectro... ) and the PDF skill helped make it much less awful today 🙂
November 10, 2025 at 11:30 AM
💻 How I use @raycast to drive a lot of work with AI tooling FAST! 🚀

When working with the best AI tools, productivity improvements are often constrained by the human operator's ability to work fast. This is one solution that works for me.

www.youtube.com/watch?v=8LB...
Raycast for Hyperactive AI Obsessives
How I use Raycast to drive a lot of work with AI tooling FAST!See also:- https://ai.intellectronica.net/- https://x.com/intellectronica- https://www.linkedin...
www.youtube.com
November 10, 2025 at 10:12 AM
Do people still use Claude Opus? What for? How?
November 9, 2025 at 9:27 AM
Curbing Claude's Enthusiasm
November 8, 2025 at 12:32 PM
Qoder CLI + Quest Mode: autonomous AI coding that goes beyond chat. Free lighting lesson with Hang Yu - Product Lead for Qoder @ Alibaba - on Qoder and when and how to delegate full tasks to AI. maven.com/p/56bd6c/au...
Autonomous AI Coding Workflows with Qoder CLI and Quest Mode
Most AI coding tools work as co-pilots. Qoder CLI and Quest mode take a different approach: autonomous coding where you delegate tasks to AI agents that execute in the background. I'll demo the workflow, explain when autonomous beats co-pilot, and share real production use cases.
maven.com
November 7, 2025 at 2:58 PM
Claude DR is the best one so far.
November 6, 2025 at 9:54 PM
The people we follow and try to learn from can be divided into two categories:
1. People who tell us "this is more complicated than you think"
2. People who tell us "this is simpler than you think"

I am only interested in (2) and I'm doing my best to be part of that category.
November 6, 2025 at 7:21 AM
Finally got around to creating my own lovely Claude Code custom status line.

directory ・ git ・ mcp servers ・ cost (if API key) ・ model (and thinking)・ context window ・ version
gist.github.com/intellectro...
My Claude Code custom status line (directory ・ git ・ mcp servers ・ cost (if API key) ・ model (and thinking)・ context window ・ version)
My Claude Code custom status line (directory ・ git ・ mcp servers ・ cost (if API key) ・ model (and thinking)・ context window ・ version) - statusline.png
gist.github.com
November 5, 2025 at 11:31 AM
The recipe is AI-generated, the photo and sandwich are real (according to @hetty).
November 5, 2025 at 11:16 AM
Congrats to @warpdotdev on moving all the way to metered token consumption (with BYOK, on top of a base subscription for the excellent product). The all-you-can-eat model was never sustainable and I am thrilled that the industry is moving away from it.
November 4, 2025 at 10:44 PM
Thank you @AnthropicAI 😍 No idea how I can spend $1,000 in two weeks but ... challenge accepted 🫡
November 4, 2025 at 5:14 PM
Nice that MSEdit is finally available in Homebrew. Very nice to have for a quick edit job in the terminal. formulae.brew.sh/formula/msedit
msedit
Homebrew’s package index
formulae.brew.sh
November 4, 2025 at 1:40 PM
Is it silly to run Codex from within Claude? I love working in Claude Code and the Claude 4.5 models are great for almost anything, but when it comes to thorough analysis and planning, I still prefer GPT-5-high, so this is how I run it. gist.github.com/intellectro...
Claude Code /command to run Codex CLI - Great for complex planning and analysis
Claude Code /command to run Codex CLI - Great for complex planning and analysis - codex.md
gist.github.com
November 3, 2025 at 9:36 PM
An ability to change my stubborn mind is that tool that has so far served me best in coping with rapid technological change.

A year ago I thought that agentic AI coding will never work. Now it's most of what I do (incl. teaching the #1 course about it).
November 3, 2025 at 6:21 PM
Struggling with terminal-only AI coding agents? Vibe Kanban revolutionizes it with a web-based Kanban board for managing parallel CLI tools like Claude Code & Codex. Track tasks, execute in isolation, review diffs — effortless! 🚀 elite-ai-assisted-coding.dev/p/vibe-kanb...
Vibe Kanban Tool Review
A local web-based interface for managing parallel CLI coding agents
elite-ai-assisted-coding.dev
November 3, 2025 at 4:15 PM
Cool trick for running Vibe Kanban from the Dock, or Spotlight, or Raycast:
1. Create a script like:
```
#!/bin/bash

osascript <<EOF
tell application "Terminal"
do script "npx -y vibe-kanban; exit"
activate
end tell
EOF
```
November 2, 2025 at 9:23 PM