HNHacker News
TopNewBestAskShowJobs

ewoodrich

2,349 karma · joined October 23, 2011

submissionscomments
ewoodrich··on Run interactive commands in Gemini CLI
I stopped reading at that point, it was a signal that I’d just be reading another several paragraphs of repetitive prose with random bolded text. It also put such strange over emphasis on an implementation detail that is pretty much irrelevant to users which made it actively distracting on top of being an obvious LLMism.
ewoodrich··on Bay Area tech CEO says test project likely struck United flight at 36,000 feet
It made more sense after seeing one of the pics posted of the ballast that is basically a bag of sand.

https://xcancel.com/DJSnM/status/1980433951551533203#m

ewoodrich··on Starcloud
> Sam Altman of Loopt is one of the most successful alumni, so we asked him what question we could put on the Y Combinator application that would help us discover more people like him. He said to ask about a time when they'd hacked something to their advantage—hacked in the sense of beating the system, not breaking into computers. It has become one of the questions we pay most attention to when judging applications.
ewoodrich··on LLMs can get "brain rot"
Lately the Claude-ism that drives me even more insane is "Perfect!".

Particularly when it's in response to pointing out a big screw up that needs correcting and CC utterly unfazed just merrily continues on like I praised it.

"You have fundamentally misunderstood the problems with the layout, before attempting another fix, think deeply and re-read the example text in the PLAN.md line by line and compare with each line in the generated output to identify the out of order items in the list."

"Perfect!...."

ewoodrich··on ChatGPT Atlas
When ChatGPT was first released I quickly lost count of the number of HN comments predicting Google had about 6-12 months before their search monopoly and ad revenue evaporated.
ewoodrich··on Claude Code on the web
Right, that is closer to what I was hoping this announcement would be. I really just want a (mobile/web) companion to whatever CLI environment I have Claude Code running in. That would perfectly fill in the exact niche missing in my local dev server VM setup I remote into with any combination of SSH, VS Code Remote, or via Web (VS Code Tunnel from vscode.dev and a ttyd remote CLI session in the browser).

It would be great to be able to check in on Claude on a walk or something to make sure it hasn't gone off the rails or send it a quick "LGTM" to keep moving down a large PLAN.md file without being tethered to a keyboard and monitor. I can SSH from my phone but the CLI ergonomics are ... not great with an on screen keyboard, when all it really needs is just needs a simple threaded chat UI.

I've seen a couple Github projects and "Happy Coder" on a Show HN which I haven't got around to setting up yet which seem in the ballpark of what I want, but a first party integration would always be cool.

ewoodrich··on Claude Code on the web
I've been dealing with this on Codex a lot lately. It confidently wraps up a task, I go to check it's work... and it's not even close.

Then I do a double take and re-read the summary message and realize that it pulled a "and then draw the rest of the owl", seemingly arbitrarily picking and choosing what it felt like doing in that session and what it punted over to "next steps to actually get it running".

Claude is more prone to occasional "cheating" with mocked data or "tbd: make this an actual conditional instead of hardcoded If True" stuff when it gets overwhelmed which is annoying and bad. But it at least has strong task adherence for the user's prompt and doesn't make me write a lawyer-esque contract to avoid any loopholes Codex will use to avoid doing work.

ewoodrich··on Comparing the power consumption of a 30 year old refrigerator to a new one
Same. Sometimes I'd be sitting in a different room not exactly "hearing" anything except a vague sense of heightened perception and unease, and eventually decide to get up to check the TV room on the other side of the house and sure enough, someone left the TV on. The feeling of utter blissful silence afterwards in comparison was incredibly relaxing though, I do kinda miss that part...
ewoodrich··on Claude Skills
Right, but typically making those kind of mistakes creates more work for yourself and with the benefit of experience you get better at recognizing the red flags to avoid getting in that situation again. but it

Which is why I think the parent post had a great observation about human problem solving having evolved in a universe inherently formed by the additive effect of every previous decision you've ever made made in your life.

There's a lot of variance in humans, sure, but inescapable stakes/skin in the game from an instinctual understanding that you can't just revert to a previous checkpoint any time you screw up. That world model of decisions and consequences helps ground abstract problem solving ability with a healthy amount of risk aversion and caution that LLMs lack.

ewoodrich··on Gemini 3.0 spotted in the wild through A/B testing
I still don’t really understand the criticism of AI Studio, it’s just the developer environment for trying out models with super low barrier to entry.

Either with the web UI a la OpenAI Playground where you can see all the knobs and buttons the model offers, or by generating an API Key with a couple clicks that you can just copy paste into a Python script or whatever.

It would be much less convenient if they abandoned it and forced you to work in the dense Google Cloud jungle with IAM etc for the sake of forced “simplicity” of offering models in one place.

ewoodrich··on Claude Skills
Whenever I watch Claude Code or Codex get stuck trying to force a square peg into a round hole and failing over and over it makes me wish that they could feel the creeping sense of uncertainty and dread a human would in that situation after failure after failure.

Which eventually forces you to take a step back and start questioning basic assumptions until (hopefully) you get a spark of realization of the flaws in your original plan, and then recalibrate based on that new understanding and tackle it totally differently.

But instead I watch Claude struggling to find a directory it expects to see and running random npm commands until it comes to the conclusion that, somehow, node_modules was corrupted mysteriously and therefore it needs to wipe everything node related and manually rebuild the project config by vague memory.

Because no big deal, if it’s wrong it’s the human's problem to untangle and Anthropic gets paid either way so why not try?

ewoodrich··on Two things LLM coding agents are still bad at
Hah I also happened to use Claude recently to write basic MQTT code to expose some data on a couple Orange Pis I wanted to view in Home Assistant. And it one-shot this super cool mini Python MQTT client I could drop wherever I needed it which was amazing having never worked with MQTT in Python before.

I made some charts/dashboards in HA and was watching it in the background for a few minutes and then realized that none of the data was changing, at all.

So I went and looked at the code and the entire block that was supposed to pull the data from the device was just a stub generating test data based on my exact mock up of what I wanted the data it generated to look like.

Claude was like, “That’s exactly right, it’s a stub so you can replace it with the real data easily, let me know if you need help with that!” And to its credit, it did fix it to use actual data but I re-read my original prompt was somewhat baffling to think it could have been interpreted as wanting fake data given I explicitly asked it to use real data from the device.

ewoodrich··on Two things LLM coding agents are still bad at
> When the agents needs to run commands like 'npm test' in a sub-directory, they almost never get it right the first time)

I was running into this constantly on one project with a repo split between a Vite/React front end and .NET backend (with well documented structure). It would sometimes go into panic mode after some npm command didn’t work repeatedly and do all sorts of pointless troubleshooting over and over, sometimes veering into destructive attempts to rebuild whatever it thought was missing/broken.

I kept trying to rewrite the section in CLAUDE.md to effectively instruct it to always first check the current directory to verify it was in the correct $CLIENT or $SERVER directory. But it would still sometimes forget randomly which was aggravating.

I ended up creating some aliases like “run-dev server restart” “run-dev client npm install” for common operations on both server/client that worked in any directory. Then added the base dotnet/npm/etc commands to the deny list which forced its thinking to go “Hmm it looks like I’m not allowed to run npm, so I’ll review the project instructions. I see, I can use the ‘run-dev’ helper to do $NPM_COMMAND…”

It’s been working pretty reliably now but definitely wasted a lot of time with a lot of aggravation getting to that solution.

ewoodrich··on Vibe engineering
I was writing up some recommendations to my team for using Github Copilot in its different modes + Roo Code/Cline and tried so hard to avoid using “vibe coding” in a professional context. For sake of clarity I ended up including it once in a “you may have heard the term…” style reference but it just comes across as so deeply unserious and felt ridiculous being forced to use it at all.

I also landed on “agentic coding” FWIW. It’s a bit clunky but didn’t feel like an embarrassment to the entire software industry typing it so I’ll happily use it over the alternative.

ewoodrich··on Claude Sonnet 4.5
But then that goes back to the original question, considering my own experiences observing the amount of damage CC or Codex can do in a working code base with a couple tiny initial mistakes or confusion about intent while being left unattended for ten minutes, let alone 30 hours....
ewoodrich··on Claude Sonnet 4.5
An agent being tasked to resolve simple issues from a compiler/test suite/linter/etc is pretty typical use case. It's not clear in this example if the linter was capable of auto fixing the problem, so ordinarily this would be a case where you'd hope an LLM would shine given specific, accurate context and known solution.
ewoodrich··on Claude Sonnet 4.5
Gemini 2.5 Pro seems to have a tic where after an initial failed task, it then starts asserting escalating levels of confidence for each subsequent attempt. Like it's ever conscious of its failure lingering in its context and feels the need to over compensate as a form of reassuring both the user and itself that it's not going to immediately faceplant again.
ewoodrich··on To AI or not to AI
They mean it doesn't learn from experience/mistakes after spending time on your codebase. There are workarounds like documenting in CLAUDE.md/CODEX.md/.roorules which dump it back into the active context but are hit and miss in my experience. It's definitely better than nothing but Claude still routinely ignores important directives whenever it's in the right mood.
ewoodrich··on The AI coding trap
Hah today I discovered Claude Code has been copy/pasting gigantic blocks of conditions and styling every time I ask it to add a "--new" flag or whatever in a once tiny now gigantic script I've been adding features to.

It worked fine until recently where I will ask it to tweak some behavior of a command with a flag and it does a diff with like hundreds of lines. So now it's struggling to catch every place it needs to change some hardcoded duplicate values it decided to copy/paste into two dozen random places in the code.

To be fair it is doing a decent job unfucking it now that I noticed and started explicitly showing it how ridiculously cumbersome and unmaintainable it made things with specific examples and refactoring. But if I hadn't bothered to finally sit down and read through it thoroughly it would have just become more broken and inconsistent as it grew exponentially.

ewoodrich··on Improved Gemini 2.5 Flash and Flash-Lite
Yep Kilo (and Cline/Roo more recently) push these free trial of the week models really hard, partially as incentive to register an account with their cloud offering. I began using Cline and Roo before "cloud" features were even a thing and still haven't bothered to register, but I do play with the free Kilo models when I see them since I'm already signed in (they got me with some kind of register and spend $5 to get $X model credits deal) and hey, it's free (I really don't care about my random personal projects being used for training).

If xAI in particular is in the mood to light cash on fire promoting their new model, you'll see it everywhere during the promo period, so not surprised that heavily boosts xAI stats. The mystery codename models of the week are a bit easier to miss.

ewoodrich··on Yt-dlp: Upcoming new requirements for YouTube downloads
If it works like Youtube TV you are given the option to switch household locations when you get the nag screen.
ewoodrich··on Snapdragon X2 Elite ARM Laptop CPU
Yes, looking at the FAQ, for example, it claims that USB is flat out unsupported on Linux which hasn't been true for 4+ years so it's very outdated.
ewoodrich··on Snapdragon X2 Elite ARM Laptop CPU
I've used VS Code on ChromeOS with the GPU acceleration flag for many, many years without any issues on a couple different devices (x64 and more recently, arm64). It can even hide the window chrome so looks 1:1 with VS Code on any other platform. And many other GUI Linux apps where the Android version feels too much like a toy in comparison, it's an incredibly versatile feature for dev work.
ewoodrich··on MrBeast Failed to Disclose Ads and Improperly Collected Children's Data
Trisolarans did evolve under much different selection pressure that requires effective mass coordination to dehydrate when deemed necessary for civilization's survival. That would probably select for rigid adherence to rules of communication vs critically evaluating every interaction. Or else you'd get a lot of "But am I being tricked into dehydrating to steal my belongings and take my job at the sophon factory? No thanks, I'll pass."
ewoodrich··on AI was supposed to help juniors shine. Why does it mostly make seniors stronger?
I have been burned so many times asking LLMs about whether some tool/app/webapp has a feature, if so where I can find or enable or disable it, etc. The number of "just plausible enough to believe" hallucinations I've got back as answers is absolutely maddening.

I've lost count of how many times I've asked whether some command line tool has an option or config available for some niche case and ChatGPT or Gemini shouts "Yes! Absolutely! just use '--use-external-mode' to get the behavior you want, it's that simple!" and it's 100% hallucination created by mangling together my intent with a real option in the docs but which in reality does not actually exist nor has it ever existed. It's even worse with GUI/menu navigation questions I'm guessing because it's even less grounded by text-based docs and trivially easy to bullshit that an option is buried in Preferences, the External tab maybe, somewhere, probably.

The desperate personality tuning to please the user at all costs combined with LLMs inherently fuzzy averaging of reality produces negative value whenever I truly need a binary yes/no "Does X exist in Y or not?" answer to a technical question. Then I waste a bunch of time falling back to Google trying to definitively prove or disprove whether "--use-external-mode" is a real thing and sure enough, it's not.

It does occasionally lead to hilariously absurd exchanges where when challenged instead of admitting its mistake the LLM goes on to invent an elaborate entirely fabricated backstory about the implementation of the "--use-external-mode" command to explain why despite appearing to not exist, it actually does but due to conflicts with X and Y it isn't supported on my environment, etc, etc.

I use Claude Code, Roo Code, Codex and Gemini CLI constantly so I'm no kneejerk LLM hater to be clear. But for all the talk about being "a better version of Google" I have had so much of my time wasted by sending me down endless rabbit holes where I ignored my sneaking suspicion I was being lied to because the answer sounded just so plausibly perfect. I've had the most success by far as a code generation tool vs. a Google replacement.

ewoodrich··on AI was supposed to help juniors shine. Why does it mostly make seniors stronger?
This can go both ways though. I've had instances where after a long somewhat tedious back and forth with Claude Code what I'm looking for finally "clicks" and it starts one-shotting requests left and right. But then after coming up to the end of the context window, I compact or ask it to update CLAUDE.md with what it learned and it loses the nuanced understanding it built up through trial and error and starts screwing up in the exact same ways again with fresh context.

And then I waste 20 minutes bashing my head against the wall trying to write three paragraphs meticulously documenting all the key gotchas and lessons from the "good" context window with the magic combination of words needed to get Claude back in the right head space to one-shot again.

At least if I pull that off I can usually ask Claude to use it as documentation for the project CLAUDE.md with a pretty good success rate. But man it leaves a bad taste in my mouth and makes me question whether I'm actually saving time or well into sunk cost territory...

ewoodrich··on iFixit iPhone Air teardown
The fact that seemingly 90% of stolen iPhones end up pinging from Shenzhen months later seems to indicate it's the inevitable top of the pyramid for the vast majority of the stolen phone market. As you mention due to the abundance of highly sophisticated large scale reverse engineering and disassembly capability in a cat and mouse game with Apple's expanding parts serialization. Plus, maybe 1/100 victims fall prey to the various phishing or threatening harassment strategies also associated with Shehnzen pings to remove the iCloud lock giving them full resale value occasionally as additional incentive.

So despite the meme that stolen iPhones are "worthless" while locked, the presence of large buyers with demand for unlimited quantities of locked iPhones means that the fence value at the lowest level of the food chain is still more than enough to incentivize rampant theft (even if it were only like $50 to a thief for a $1200 phone, that is worth $200 to a Shehnzen buyer).

I'm making the numbers up, but the point is that even if iCloud activation lock and serialization destroys the vast majority of the phone's value it's not enough to truly discourage theft rings selling to wholesale buyers overseas. It just put the humble local neighborhood thief or opportunist reselling on eBay out of business, with the vacuum quickly filled by organized crime.

In a world where most people don't carry around cash or valuables that's the best a street level thief is going to do unless you run into the odd person with a Rolex or jewelry with rare actual pawnable value, and it's a bottomless crime of opportunity in a big city or festival/concert/etc so adds up quickly.

ewoodrich··on PYREX vs. pyrex: What's the difference?
I don't think it changes the rate of cooling that much, just more of the the heat is transferred by convection to the surrounding air instead of conduction to the surface (which for a stone countertop has a lot of mass to absorb heat making them notorious for blowing up hot glassware on contact).
ewoodrich··on The quality of AI-assisted software depends on unit of work management
I just got burned by that, it compacted and then immediately dropped what it was in the middle of working on to redo something it had already finished half an hour earlier. Which, predictably, sent it into "systematically destroy the entire working codebase" mode because the code it was now reading didn't match expectations of the original instructions. So it started extrapolating like "huh, it looks like the function already exists, therefore, user must have meant [bizarre completely out of left field guess] instead" in an escalating loop of confusion and code mangling.
ewoodrich··on Hosting a website on a disposable vape
Yep, every so often I remember staring out the window on the highway in the car as a kid and seeing single use bags snagged on fences or trees pretty much anywhere inaccessible or not routinely cleaned.

I also catch it on B roll footage in movies or shows from the 90s/2000s a lot. It’s a specific type of visual blight I rarely ever see after those ultra flimsy single use bags that could be carried dozens of miles on a gentle breeze were eliminated.

(This children's book was written basically at the tail end of the era where seeing a bag flying could conjure the imagination)

https://www.amazon.com/Bag-Wind-Ted-Kooser/dp/0763630012#ave...

← PreviousPage 5 of 24Next →