Frieda Huang

Difference for the sake of it in everything. —James Dyson

[Frieda Huang @JYFHuang]

Life is good!!!

[Sebastian Raschka @rasbt]

After 18 months of writing, coding, and experimenting, Build a Reasoning Model (From Scratch) is
finally out!

My first copies just arrived! 📚

440 full-color pages. Inference scaling, reinforcement learning, and distillation from scratch.

[Marc Andreessen 🇺🇸 @pmarca]

Learn to ship.

[nader dabit @dabit3]

agreed, and it seems to be happening much faster for companies / teams we work with including ourselves @cognition where we're at ~85% cloud, vs indie devs, which makes sense as cloud agents are much more collaborative in nature

I wrote this up when I joined with my thoughts
https://t.co/oOJhfTU3hz

[Frieda Huang @JYFHuang]

@BennyKokMusic Would be nice to try this as well: your agent talks to my agent.

[Frieda Huang @JYFHuang]

Openclaw + whoop mcp to monitor my health!!!

[Frieda Huang @JYFHuang]

@evisdrenova Honestly, if we think about it -- it's similar to talking to a human. i.e., when talking to a friend we know who share some of experiences, we "assume" they remember the conversations or things that happened in the past. You pull / create "chats" through a human interface.

[Evis Drenova @evisdrenova]

my ideal AI interface is a single never-ending chat thread.

i don't want to think about the concepts of sessions, context windows, worktrees, mcp servers or really anything else.

The harness should automate everything transparently.

[Frieda Huang @JYFHuang]

@rasbt Wowww!!! Was just thinking about switching to local agents!!! Codex + Claude Code getting expensive...

[Sebastian Raschka @rasbt]

I put together a new article on setting up local coding agents with open-weight models. Everything runs 100% locally.

I thought it might be useful putting this together because many people asked me about my setup in the past, and I thought it would also motivate people to get started tinkering with local models for serious work (yes, things got incredibly capable this year with better LLMs and better harnesses).

So, here's a walkthrough of how to connect a local LLM to a local coding harness (could be Claude Code or Codex, which you may already be familiar with).

I also included some assessment notes that are useful as a checklist to select between and consider certain LLMs over others:

- Checking RAM usage at long contexts to see if the model is suitable for real work
- Measuring prefill and decoding tok/sec to see whether it's fast enough to not be annoying
- Making sure the model has sufficient tool-calling capabilities in theory
- Assessing whether the model can solve some more challenging tasks when used in a coding harness.

Of course, there are always more specialized tools that can squeeze a bit more performance out of things, but I hope this is a good starter kit that stays flexible; that is you can easily switch to newer models as they are released or even tap into cloud models in your familiar harness if the current ones are not sufficient enough for a given task.

[Chris Powers @fortworthchris]

The way someone uses AI tells you a lot about them.

"Mark Cuban (@mcuban) has a quote I wish were mine. Some people use AI so they can learn everything, and some use AI so they don't have to learn anything."

Chris Koerner (@mhp_guy) is firmly in the first camp. And he says he can spot the people who aren't.

"At my church I hear talks now and I can tell. That's Opus 4.7. And I think that's a shame."

His own rule is simple. AI shouldn't save you time. It should make the work better.

"If I have three hours to build a lesson, I'm still going to spend three hours. I'll use AI the whole time, and it'll be far better. I could be done in three minutes. But I don't."

In this clip, Chris Koerner on using AI to raise your standard instead of lowering it.

[Peter Steinberger 🦞 @steipete]

I feel the same about shared office spaces.

[Frieda Huang @JYFHuang]

@paulg @ycombinator Isn’t that the coding agent we use everyday?!

[François Chollet @fchollet]

The hardest problems are rarely solved by adding more complexity to the solution -- they are solved by reframing the question until a simpler, clearer answer reveals itself.

[jack @jack]

we're going to start talking a lot more about our intelligence tools. this is the beginning of the beginning.

[Charlie Marsh @charliermarsh]

At OpenAI, we're continuing to bet on Rust as the future of systems programming.

I'm proud to announce that we're making a $600,000 commitment to the Rust Foundation, which combines our Platinum membership with additional support for maintainer efforts across the Rust ecosystem.

[Frieda Huang @JYFHuang]

2 things I’m fully convinced of:

1. Anything that can be automated should be automated.
2. Build a machine that build machines.

[Frieda Huang @JYFHuang]

funniest thing I heard this week: birds will fly, fish will swim, and LLMs will slop.

[Frieda Huang @JYFHuang]

The limits of my language mean the limits of my world. – Ludwig Wittgenstein

The limits of my tokens mean the limits of my productivity. - me

[Frieda Huang @JYFHuang]

Beautiful day!!

[Frieda Huang @JYFHuang]

Less is more.