Jon Sykes

Coder + founder • AI Enthusiast • Always building • No-hype tech takes

[Jon Sykes @JonCSykes]

Last week I asked the question:

Can Qwen 3 Coder Next actually ship a real app from spec to prototype running fully local?

I put it to the test with Opencode CLI on a Mac Studio M3 Ultra.

It created this entire prompt auditing tool, super impressed:

[Jon Sykes @JonCSykes]

@OfficialxKee Michelle Branch is fire.

[Jon Sykes @JonCSykes]

@OfficialxKee Voiced by Matt Mercer, a legend.

[Jon Sykes @JonCSykes]

@OfficialxKee @shirtsthtgohard 100

[Yuchen Jin @Yuchenj_UW]

Finally, this is happening.

[Jon Sykes @JonCSykes]

I think plugins are going to be powerful as well, but native chat should be considered. Keep the plugins for integrations.

Thank you for attending my Ted Talk.

[Jon Sykes @JonCSykes]

I spent last night playing around with Paperclip and I think it is so close to being what I was hoping to get out of Openclaw.

@dotta I know you don't want chat as a core feature, but a native way to chat with your "employees" would put this thing on the next level.

[OpenAI @OpenAI]

5.4 sooner than you Think.

[Jon Sykes @JonCSykes]

@tejgw Chronicle

[Jon Sykes @JonCSykes]

Considering a writeup for new grads and junior devs on how to stand out when everyone has AI.

Distributed systems, application architecture, frameworks, why understanding how things work > generating code.

Would you find that useful?

[Jon Sykes @JonCSykes]

Are you using AI to automate your integration testing?

If so, how are you actually using it? Curious what's working in the wild.

[Jon Sykes @JonCSykes]

@OfficialxKee This looks like Figma on crack. You have to stop giving me more things that destroy my weekends. 😅

[Jon Sykes @JonCSykes]

@AnthropicAI Thank you for doing the right thing.

[Jon Sykes @JonCSykes]

I love living vicariously through your experiments. I need to reduce the number of plates I have spinning so I can do cool stuff like this.

[Andrej Karpathy @karpathy]

I had the same thought so I've been playing with it in nanochat. E.g. here's 8 agents (4 claude, 4 codex), with 1 GPU each running nanochat experiments (trying to delete logit softcap without regression). The TLDR is that it doesn't work and it's a mess... but it's still very pretty to look at :)

I tried a few setups: 8 independent solo researchers, 1 chief scientist giving work to 8 junior researchers, etc. Each research program is a git branch, each scientist forks it into a feature branch, git worktrees for isolation, simple files for comms, skip Docker/VMs for simplicity atm (I find that instructions are enough to prevent interference). Research org runs in tmux window grids of interactive sessions (like Teams) so that it's pretty to look at, see their individual work, and "take over" if needed, i.e. no -p.

But ok the reason it doesn't work so far is that the agents' ideas are just pretty bad out of the box, even at highest intelligence. They don't think carefully though experiment design, they run a bit non-sensical variations, they don't create strong baselines and ablate things properly, they don't carefully control for runtime or flops. (just as an example, an agent yesterday "discovered" that increasing the hidden size of the network improves the validation loss, which is a totally spurious result given that a bigger network will have a lower validation loss in the infinite data regime, but then it also trains for a lot longer, it's not clear why I had to come in to point that out). They are very good at implementing any given well-scoped and described idea but they don't creatively generate them.

But the goal is that you are now programming an organization (e.g. a "research org") and its individual agents, so the "source code" is the collection of prompts, skills, tools, etc. and processes that make it up. E.g. a daily standup in the morning is now part of the "org code". And optimizing nanochat pretraining is just one of the many tasks (almost like an eval). Then - given an arbitrary task, how quickly does your research org generate progress on it?

[Jon Sykes @JonCSykes]

@Tailscale @lmstudio Well there goes my night...

[Tailscale @Tailscale]

Use local models on remote devices you control—as if they were local.

- Introducing LM Link from @LMStudio.
- Encrypted, identity-based access to your own LLM hardware.
- No public endpoints. No API key sprawl.

https://t.co/SGJC0JhlbU

[Jon Sykes @JonCSykes]

@gregisenberg @startupideaspod Heaper dropped last week, looks really promising.

[Jon Sykes @JonCSykes]

@github should start building out a native solution for this. Even if it is only available for Enterprise plans.

[Peter Steinberger 🦞 @steipete]

Been wrangling a lot of time how to deal with the onslaught of PRs, none of the solutions that are out there seem made for our scale.

I spun up 50 codex in parallel, let them analyze the PR and generate a JSON report with various signals, comparing with vision, intent (much higher signal than any of the text), risk and various other signals.

Then I can ingest all reports into one session and run AI queries/de-dupe/auto-close/merge as needed on it.

Same for Issues. P rompt R equests really are just issues with additional metadata.

Don't even need a vector db. Was thinking way too complex for a while.

There's like 8 PRs for auto-update in the last 2 days alone (still need to ingest 3k PRs, only have 1k so far).

[Jon Sykes @JonCSykes]

@albertadevs You wore it better, and with the utmost authority.

[Jon Sykes @JonCSykes]

Here is the link to @janlunge launch video for Heaper:
https://t.co/TTyA1kKtY2

[Jon Sykes @JonCSykes]

Just saw the Heaper launch video on yt. @janlunge this looks amazing, I would promote it but I can't find the socials for it.

Have you thought about this being a great file management system for @openclaw ?

I will be investigating this further... Stay tuned!