HNHacker News
TopNewBestAskShowJobs

elliotbnvl

673 karma · joined December 3, 2019

Father, writer, engineer at NVIDIA. Opinions are my own.

https://elliotbonneville.com, https://x.com/elliotbnvl

submissionscomments
elliotbnvl··on Please Use AI
Though I deeply agree with this sentiment, the author fails to address that there can be multiple goals to an action. And writing can be for art or communication or both but not always both, and removing the art from communication doesn’t destroy its ability to communicate.
elliotbnvl··on I Am Retiring from Tech to Live Offline
This resonates with me as well. For more reasons than one: with the rise of AI (Mythos is but a pale forerunner) digital security — and by extension, digital privacy — has ceased to exist. The bomber will always win. The only way to win is not to play.
elliotbnvl··on Claude for Small Business
This is a comically self-absorbed perspective.

Why are people making things with Claude Code if not because they’re motivated?

elliotbnvl··on Show HN: Stage CLI – An easier way of reading your AI generated changes locally
While you are not factually incorrect, my expectations were subverted in the same way that OOP's were.
elliotbnvl··on Grief and the AI split
Strong agree. Needs another pass or two at editing though, some painful LLM-os sticking out there :'(
elliotbnvl··on Claude now creates interactive charts, diagrams and visualizations
And the people that still get excited about life!
elliotbnvl··on Claude now creates interactive charts, diagrams and visualizations
It's a usability / quality of life feature to me. Nothing to do with increasing perceived confidence. I guess it depends on how much you already (dis)trust LLMs.

I'm finding more and more often the limiting factor isn't the LLM, it's my intuition. This goes a way towards helping with that.

elliotbnvl··on GPT-5.4
Looks like it's an order of magnitude off. Missprint?
elliotbnvl··on Enable CORS for Your Blog
This is something Opus 4.6 likes to generate a LOT for some reason.
elliotbnvl··on Text Your Site: Realtime multiplayer LLM-powered text-to-website demo
Hi all, author here. Please don't go to this website. It's expensive for me. It's a realtime multiplayer text-to-website chat, powered by Sonnet 4.6, that I built to showcase the incredible agency and speed frontier models have attained. I am paying API costs out-of pocket. Call it a business expense?

It's inspired by Pieter Levels' tweet about chatting with his websites. [1]

It uses the same tool under the hood as the core web offering, bareclaw [2], a thin Claude Code CLI wrapper I created to enable custom CC client creation while leveraging your Claude Max subscription.

[1] https://x.com/levelsio/status/2023960543959101938 [2] https://github.com/elliotbonneville/bareclaw

elliotbnvl··on Following 35% growth, solar has passed hydro on US grid
I have a goal of setting up solar on my property in the woods that goes directly to a wall of batteries, maybe Tesla, maybe something else. But definitely not going back into the grid. Does anybody have suggestions or advice on how to do this?

Who are the best companies doing this right now in New England? What products are folks using to store electricity? Are there any good resources for this kind of thing?

elliotbnvl··on Claude Code Remote Control
Happy you liked it! Always really nice to get positive feedback.

I’ll have to check out the journaling article. I’ve been journaling a lot more lately!

elliotbnvl··on Claude Code Remote Control
Wispr Flow on mobile fills this gap.
elliotbnvl··on Claude Code Remote Control
I’ve been doing this for a while [1], but ultimately settled on a building a thin transport layer for Telegram to accept and return media, and persistent channels, vastly improved messaging UX, etc. and ended up turning this into a ‘claw with a heartbeat and SOUL [2].

[1] https://elliotbonneville.com/phone-to-mac-persistent-termina...

[2] https://elliotbonneville.com/claude-code-is-all-you-need/

elliotbnvl··on Bareclaw: Claude Code Is All You Need
Hey cool, glad you are enjoying! I am having fun with it also :)
elliotbnvl··on Bareclaw: Claude Code Is All You Need
Haha. Just following the rules ;)
elliotbnvl··on Bareclaw: Claude Code Is All You Need
Amidst all the ‘claw hype, folks have lost sight of the fact that we already have a high quality, trusted coding agent that with a few lines of code can become proactive, recursively self improving, and is supported by a large dedicated engineering organization who are commercially incentivized to improve it.

Claude Code is all you need.

Well, that and a persistent transport layer that exposes a messaging endpoint you can text from your phone. And that’s bareclaw: just ~1200 LOC of TypeScript and you’re off to the races on the Claude Max plan with no OAuth token theft hijinks.

elliotbnvl··on I hate AI side projects
What kinds of other distribution strategies are you thinking of?
elliotbnvl··on Cothought: A Markdown zettelkasten journal for your life, in Claude Code
A rash of these second mind type tools have emerged, but they all seem overly focused (business or whatever), or performative.

This one started from me just typing stuff into Claude Code, then adding more ideas and skills / formatting changes as I went until it grew into something useful enough I started telling other people about what it's done for me. Then they asked me to share it so I packaged it up, polished it a bit, and here we are! There's no other editor, I don't use Obsidian or anything with it (although you definitely could). I just find the simplicity of one interface to be really freeing and helps me focus on actually journaling instead of maintaining a card / link system.

It helps me think and reflect across time in a way that my meat mind seems to be not as good at. It's just so fast at searching through things and making associations even with the default Opus 4.6 somewhat limited context. I've ended up capturing and working on so many more of my ideas and I'm so much more inspired on a day to day basis. I hope you enjoy it as much as I do.

It's all on GitHub here, and obviously free and open source: https://github.com/elliotbonneville/claude-cothought

elliotbnvl··on How to stop being boring
Love the article. My fav quote:

> The things on your cringe list are probably the most interesting things about you.

elliotbnvl··on I found a useful Git one liner buried in leaked CIA developer docs
The deps question is huge, let me know if you solve it.
elliotbnvl··on The path to ubiquitous AI (17k tokens/sec)
That… what…
elliotbnvl··on Cothought: Claude as text editor, thinking journal
I’ve been using this system for a while to great effect, and a couple of folks have asked me to share it, so I bundled it up with a pretty home page. It’s kind of a paradigm shift from existing journaling tools and uses of LLMs for writing and thinking, but very inline with it as a code editing agent.
elliotbnvl··on AI makes you boring
I was onboard with the author until this paragraph:

> AI models are extremely bad at original thinking, so any thinking that is offloaded to a LLM is as a result usually not very original, even if they’re very good at treating your inputs to the discussion as amazing genius level insights.

The author comes off as dismissive of the potential benefits of the interactions between users and LLMs rather than open-minded. This is a degree of myopia which causes me to retroactively question the rest of his conclusions.

There's an argument to be made that rubber ducking and just having a mirror to help you navigate your thoughts is ultimately more productive and provides more useful thinking than just operating in a vacuum. LLMs are particularly good at telling you when your own ideas are un-original because they are good at doing research (and also have median of ideas already baked into their weights).

They also strawman usage of LLMs:

> The way human beings tend to have original ideas is to immerse in a problem for a long period of time, which is something that flat out doesn’t happen when LLMs do the thinking. You get shallow, surface-level ideas instead.

Who says you aren't spending time thinking about a problem with LLMs? The same users that don't spend time thinking about problems before LLMs will not spend time thinking about problems after LLMs, and the inverse is similarly true.

I think everybody is bad at original thinking, because most thinking is not original. And that's something LLMs actually help with.

elliotbnvl··on Famous Signatures Through History
That is wild. Hahaha.
elliotbnvl··on Famous Signatures Through History
Ok, got Cranach's in (plus some more "famous" signatures, about 10 more). Cranach's might be my personal favorite signature of the lot. Dragon imagery, minimalism, cool backstory.

There were no SVGs of it available online so I had to trace a PNG but Claude handled that like a champ with a combination of imagemagick and potrace:

  magick cranach-autograph.png -colorspace Gray -threshold 70% cranach.pbm
  potrace cranach.pbm -s -o cranach.svg --tight --alphamax 1.0 --opttolerance 0.5
elliotbnvl··on Famous Signatures Through History
This was surprisingly tricky, but fixed! The issue was that short tap strokes were being fed through the stroke-rendering library (perfect-freehand), which applies start/end taper over a configurable length. When the total stroke distance is shorter than the combined taper length, the stroke tapers to nothing — so dots just disappeared. The fix detects tap-like strokes and renders them directly as small ellipses, bypassing the stroke pipeline.
elliotbnvl··on Famous Signatures Through History
Ahh good one, I added this in! That is amazing.
elliotbnvl··on Famous Signatures Through History
Good catch, missed that.

edit: went through and de-LLMified it, mostly, just for you :)

elliotbnvl··on Famous Signatures Through History
The latter, but now it’s going to be the former because you just nerdsniped me with the precision of a Navy SEAL.
← PreviousPage 2 of 9Next →