HNHacker News
TopNewBestAskShowJobs

joshstrange

23,813 karma · joined July 11, 2011

Email: Josh@JoshStrange.com Github: https://github.com/joshstrange

[ my public key: https://keybase.io/joshstrange; my proof: https://keybase.io/joshstrange/sigs/9yQAUjB3P9Xt5_LeWFPzRkPZv_m7q7M-EgiD3XOB6eo ]

submissionscomments
joshstrange··on $48 e-paper display can be whatever you want it to be
I have a couple of these and they are great. I'm transitioning to BTLE e-ink screens (hand-assembled, I wish there were nice off-the-shelf versions) because of their much lower power draw.

Currently I have a handful of the all-in-on seeed studio reTerminal E1001's that rotate book quotes I've highlighted on my kindle (jailbroken, KOReader+BookOrbit) once a day so they can last months on a battery charge. But with BTLE you can go years on a single charge, update more often, and push to the device to force an update (all things not possible when deep-sleeping wifi for 23.99 hours out of the day).

e-ink displays are so nice and once I get the BTLE+OpenDisplay setup perfected I'll probably grab ~10 more boards/displays for other places around the house.

joshstrange··on Herdr has released native app
If you are on iOS, Heeler is a great Herdr client.
joshstrange··on Patrick Stewart has died
I'm finishing up my Voyager watch right now and was already considering TNG next, this seals the deal. TNG is such a cornerstone to my youth and growing up. Everyday after school I'd come home and watch TNG reruns (Tivo). Call it whatever you will but Star Trek and the ethos in it will always hold a very special place in my heart, I've always held it up as the model to aim for. He will be missed.
joshstrange··on DoorDash to acquire Instacart
> Quite a few of them spoke no English, so yes, they'd just present their phone screen. I didn't take it as rude.

I want to be clear, the ability to speak English is not what I took issue with. That doesn't bother me and I'm fine with just showing the phone screen, no conversation. The issue I had was cutting the line and essentially demanding they be helped immediately. Wait in line like everyone else. This is a small food stall in a food hall with often just 1 person working the front. If they are actively helping another customer and have a line, I find it unacceptable to think you can cut to the front and immediately demand attention. What message does that send to the customers in line? "You're less important than delivery orders", and I can tell you that isn't the case due to the large chunk that DD takes.

My family member has even considered shutting down DD orders temporarily if/when the drivers act this way and publishing on social media about it. According to them, the loss in sales (and overall profit, which is minor on DD orders) wouldn't matter compared to the publicity/engagement (I don't know, that's not my world).

joshstrange··on DoorDash to acquire Instacart
> DoorDash to Acquire Instacart to Help Local Stores Win

> to Help Local Stores Win

Any local store will tell you DD is a necessary evil they put up with but strongly dislike overall. From the % of sales they take to the incredibly rude delivery drivers, DD is not well liked.

On multiple occasions a family member's employees (or they, themselves) have had a DD delivery driver cut to the front of the line, interrupt the cashier, and shove their phone in their face. No words, no anything, just "here is the code/screen/whatever I need to show you to pick up the order". DD doesn't let stores rate delivery drivers so there isn't any recourse.

joshstrange··on MacBook Pro with PROLED XDR Display and Touch
Does anyone use/like touch on their windows (linux?) laptops? I've never really been drawn to wanting to do that. My mouse is far more accurate and doesn't block my view. But maybe it's just because I don't have it and I would find tons of uses for it. Also, I don't love the idea of putting fingerprints all over my screen. I guess I can understand for testing iOS/iPad apps but past that I just don't see myself wanting to use that.

Then again, it really doesn't matter to me at all what the screen looks like or what features it has, I use my MBP docked 99% of the time with the screen closed, to the point that my next computer will almost certainly be a Mac Studio and I'll just use my MBP on the rare occasions I need it on the go.

joshstrange··on Gemini 4 Carbon
> Oh come now, this rolled release is now the standard for frontier models.

Sort of? Yes, OpenAI has been bad about slow rolling but Anthropic's last few models have been available essentially at the same time as they announced it. In most cases I updated Claude Code and was using the new model right away.

Compare that to OpenAI who sometimes says "Available to all Plus/Pro/etc plans" without saying anything about "rolling out" and then it's a crap-shoot on when access will actually roll out. That said, they have gotten better, it used to be 1-3 weeks sometimes to get a new model/feature they announced was available for your plan.

And then there is Google... Who loves to talk about models like this and then give no timeline for release. Even when they do release the models it's miserable to find out how to even use it. Is it in the API? Which API? Google Cloud/Vertex/Gemini Cloud? Whatever it's called this week? Coding plan?

No, what Google does (or rather fails to do) in its rollouts does not compare to OpenAI/Anthropic.

As for the AA page, it's just embarrassing. Any smaller company doing that would get roasted, Google doesn't get a pass, learn how to coordinate or wait for something to actually be true/live before posting.

joshstrange··on Gemini 4 Carbon
Ha, I didn't see your comment before I posted mine [0], there were still only 2-3 comments when I opened the thread but yeah. I completely agree. I don't think they are lying, but I still don't love the "it's the best, but you can't try it".

[0] https://news.ycombinator.com/item?id=50277040

joshstrange··on Gemini 4 Carbon
> Carbon ranks #1 on the recently updated v5.0 of Artificial Analysis Intelligence Index at 59.

Clicks link

Carbon isn't anywhere on the page and you can't toggle it on either. I'm sure it's just timing and it will be there soon but still just look amateur, which is about how I feel about Google and AI overall. I know the models have improved but I still have a bad taste in my mouth from testing Gemini in the past and I still feel that Google has an overly complicated "how to use our models" that changes way too often. That coupled with "you could lose access to your google account if we don't like what you do with Gemini" means I have little desire to test it out.

On top of all of that I'm completely over announcing a new model that we can't actually play with or try out. I don't think Google is lying about how well their model scored exactly but it still has big "my girlfriend goes to another school, you don't know her"-energy.

joshstrange··on Kentucky man becomes the first humanist granted religious exemption from AI
> "And offloading that labor, regardless of whether it's taxing or not, is undercutting my ability to be a moral actor."

The objection in the article is specifically about offloading thinking to technology, which is something we've been doing for decades, or depending on your definition, since we first started using tools.

If offloading cognitive labor is the core problem, then what makes AI fundamentally different from the countless other technological advances that have allowed us to offload cognitive labor?

joshstrange··on Kentucky man becomes the first humanist granted religious exemption from AI
The car he drives to work, the house he heats, the electricity he consumes daily, all that labor is fine to offload, but AI is the line in the sand?
joshstrange··on Kentucky man becomes the first humanist granted religious exemption from AI
Where is the line? "AI" is completely arbitrary, might as well say a calculator is immoral because it offloads thinking. That my main issue with these people is that they think AI/LLMs are different than everything that came before them, yes they are more capable but all of computing has involved "offloading" effort/thought to computers. `
joshstrange··on 'I Felt Blind Panic.' They Opened Their New $2k iPhone Duo–and It Broke
There’s a good chance I’m wrong about this, but I thought they lowered the price to account for not including Apple Care? So you’re correct, it’s no longer included, but the price per month and total cost of ownership is the same?
joshstrange··on I will no longer upload to NEW
We’re still experimenting. Automatically fixing obvious issues raised by the agent (at another agent’s discretion) or presenting a list of the feedback items with fix/refute suggestions are things we’ve toyed with.

Personally I have an agent review the feedback and break down each item so it’s easier to review (helps remove some of the jargon/short-hand that agents use sometimes). Then I decide what to have an agent address and what to push back on. Like I said, I’m working on finding a way to automate part of that but haven’t settled on one I like yet.

joshstrange··on Proton to 'Re-Interview' Staff About AI to See Who Gets to Keep Their Job
>> Writing code by hand is dead.

> Is it still a controversy nowadays?

Reading HN would leave one with the conclusion that yes, it is still very controversial. We still have plenty of people who swear they will never use AI, who say it's trash, it only produces slop, that a reckoning is coming, etc.

See this post from today: https://news.ycombinator.com/item?id=50256738

joshstrange··on 'I Felt Blind Panic.' They Opened Their New $2k iPhone Duo–and It Broke
My understanding is that while it's a loan->lease change, the TCO is the same under both plans. I might agree that the Klarna version feel scummier but it doesn't change the financials for me and it's still effectively a 0% interest loan. I set aside the full cost of the phone in my budgeting app on day 1 and then draw down from that category over the span of the loan/lease so it's effectively like I'm buying it but I get to park the money in a HYSA.
joshstrange··on Ask HN: T'was the night of Claude Reset and quota left: Funniest app you built?
My go-to is adding unit tests or similar to existing projects that didn't have good coverage when originally written. It's effectively completely safe to do (only adding tests, not touching application code) and is nice to "pin" our existing functionality.
joshstrange··on 'I Felt Blind Panic.' They Opened Their New $2k iPhone Duo–and It Broke
While I don't doubt many people are doing that, using the iPhone/Apple Upgrade Program does not immediately equate to that being the case. I always use the Upgrade program but I could buy the phone outright if I wanted but why would I? It's essentially a 0% interest loan if I buy the phone out at the end. Sure if you turn the phone in after a year (or two) you are leaving money on the table (vs selling it or repurposing it) but the Upgrade programs aren't only used by people spending outside their means.
joshstrange··on I will no longer upload to NEW
Agreed. It's sad how bad GH Copilot is, and thank you for reminding me to cancel that subscription. I just realized I haven't used it in a while now.

We also settled on a custom solution for reviews. We started with reviewers that claude/codex on GH actions and those were "ok" but kind of dumb and hard to tweak (GH Actions is just miserable to debug/iterate on). We still use them occasionally (@tag them to kick off a review) but we've shifted to a piece of software that we wrote that watches GH PRs and once certain labels have been added and the ticket in our ticket tracker (not GH issues) is marked as "Ready for Review" it kicks off a review and it reviews _all_ the PRs connected to the ticket. We don't have a monorepo, yet, so a new feature might span 3-4 repos (server, client1, client2, ...) and we want that work reviewed together.

We are close to having our review code spin up a fully isolated devstack running all our, many, services so it can poke around in the UI and write custom playwright scripts to test the new functionality. We have the devstack VMs working, just gluing the last of it together, it's pretty cool to see it all come together.

joshstrange··on I will no longer upload to NEW
While I, like almost everyone, doesn't love reading LLM output, code review is an excellent use for LLMs. You know what I like less than reading LLM output?

- Slow PR review

- Nitpicks on style, not substance

- Async waiting for review or re-review after making changes, harder to stay in what passes for a "flow state" nowadays

- A lot of a ceremony around straightforward changes

We use LLM reviews at my company and it's caught things human reviews have not (as in in the early days we human-reviewed code and then did tests running the LLM afterwards and it caught things we didn't). It can review code at any time of the day and re-review it just as quickly as you can address the feedback.

I predict we (my company) are less than a year out from having a fully automated pipeline for reviews with only humans getting called in when it touches certain parts of the code. We already have a crude system for detecting certain file paths that we want a human to review if touched and we tweak that list over time as needed.

joshstrange··on “AI Skills” are worthless in 2026
Tower? This is from first-hand experience. I see it within my company. I’m not saying you can’t learn or won’t “catch up”, I’m simply pointing out that using LLMs effectively is a skill you need to hone.
joshstrange··on “AI Skills” are worthless in 2026
While I understand where this author is coming from. and agree that we’ve had a number of false starts or silly concepts like “prompt engineer”, there is definitely a skill in using LLMs, and it is a spectrum for different people.

Being completely engrossed in the new trends is probably not necessary, but keeping abreast of what’s possible and trying new things is. People who have an ignored/avoided LLMs completely are not able to just pick them up today and be as productive as people who have been learning the best ways to use them over the last couple years.

Many of the “skills” are not new, it’s the same skills you’ve been using throughout your career, just with a new tool. But any new tool requires practice and learning what it’s capable of.

I’m constantly experimenting around the edges of what LLMs are good at because it keeps changing and just because an LLM wasn’t good at X task 6 months ago doesn’t mean it still isn’t good at it today. The only way you can know is by testing and trying out things.

I work with a number of smart people, but their LLM proficiency is a wide spectrum. LLMs might/are a leveling factor, but there is still skill in getting what you want out of an LLM. This is most apparent to me when I interact with people who aren’t developers or aren’t as technically savvy. They just aren’t able to produce or get the same things out of an LLM that a developer or more technical person can.

joshstrange··on Proton to 'Re-Interview' Staff About AI to See Who Gets to Keep Their Job
I don’t really understand the surprise in the comments. It’s been an oft-repeated line that you won’t lose your job to AI, but to someone who uses AI better than you do.

Look, I know it’s a huge shift but there is no way around it, 6-12mo ago LLM usage might have been optional or up to each employee but the models and harnesses have improved to a point that not using them is just cutting off your nose to spite your face.

I’ve reviewed plenty of code pre and post-LLM and while LLM code is not perfect, neither was the human-created code (mine included, obviously). Missing/no tests, issues that are minor but you don’t want to send it back _another round_, coworkers/employees that don’t make the changes you ask for or do it in a hacky/quick way, and generally accepting code at a level under what you wish because you can’t wait for the code to be perfect.

Now the ability to write solid code is available to all your employees and so the desired skills shifts. It makes sense to re-evaluate your workforce when that happens.

If an employee/coworker wasn’t using LLMs today I’d tell them to either start or consider finding another job because they’d be dragging down the company.

Writing code by hand is dead. It just is. In a couple years this won’t be controversial, just accepted fact.

joshstrange··on New iPad mini
I've thought a lot about that actually. I have a number of hardware purchases I'd like to make but I'm delaying due to the insane prices. That said, would I give up LLMs for the prior prices? No. Prices will come down eventually once more factories come online (it's just slow to bring up a new factory/fab, especially for RAM) and LLMs are such an awesome tool that I'll take the short-term pain of higher hardware.
joshstrange··on Active incident at Dreamhost forcing password reset
I got the email about this and realized I’ve been using them for just shy of 20 years. It was one of the original places I hosted things and where I hosted projects for other people. I haven’t actively used it years but migrating the last few things off has never been a priority. This might be the push I need to re-evaluate everything hosted there and shut down or migrate it all.
joshstrange··on Today, we hacked what we'd set out to hack: the iPhone
I jailbroke every iPhone I had for many, many years. Most of the time it was silly skins/tweaks but there were a number of things that were huge QoL improvements. I remember BiteSMS and using it as a way to reply to a text directly in the notification instead of having to leave the app. This was before iOS had that feature and before multitasking, back when if you left the game you were playing the app would quit. Also had folders before stock iOS, things like Control Center, and a whole slew of little tweaks for other apps I used.

The downside to jailbreaking was often the instability but that was also back when the phones were way less powerful. Today, armed with LLMs, I would love to have root on my iPhone again. I'd love to rip out Siri and replace it with Claude Code or something like that.

joshstrange··on Claude Now Works with Google Docs, Sheets, and Slides
And it doesn't work if you are signed into multiple Google accounts :/

Today I wanted to fix some mixed formatting in a Google Doc. Ok, I'll use Gemini, it's right there and surely it's the best. Nope, fails twice "Please try again" then suggests just trash formatting. Ok, fine, I connected this to ChatGPT, let's try that. Nope, multiple rounds of "I can't see that document" despite having the plugin configured. Ok, onto Claude... Nope, can't do it (even with a connector setup) but suggests I use this. Ok, sounds good, I'll install the plugin to Google Docs, login to Claude in the new sidebar, and.... "I can't read the document". I give up.

joshstrange··on Median product "lifespan" has fallen from ~4y down to ~1.5y
LLMs have flipped the build vs buy debate on its head for me. I also have been re-evaluating existing tools and building my own custom, just-for-me versions and it's been liberating. Using off-the-shelf software increasingly feels stifling and limiting. I do wonder if the future for many SaaS companies is to provide a solid core with a nice API/MCP/Hooks/etc to build on top of so you can add features only your company cares about instead of every SaaS trying to be a one-size-fits-all.

The cool part is building smaller tools to solve your problems also allows you link those all together (MCP/CLI/API) in ways that weren't possible in the past using LLMs as a glue layer.

joshstrange··on Speeding up code review with `small-change`
I thought for sure you might be one of my coworkers based on this comment, we do pretty much the same thing with our “LLM Review” tool. It tries its best to determine what needs human eyes (mostly through file paths and the model’s own “discretion” for changes outside the “important” paths).
joshstrange··on Claude Duet
I've used Dots and Duet now (not planning to touch Muse/Grok) and while I want Duet to be better, Dots is the better product overall right now.

After setting up Duet it suggested things it could help me with. The things it suggested were "ok" but not great suggestions IMHO. I let it go to see what it would do and I hated the UI/UX within seconds.

It started "working" and dumping "Cards" in the chat of the tickets it was looking up in my ticketing system. Completely useless to me and distracting and without the context of "why" it was very annoying. Don't throw your tool use in my face, I don't care, do the work and come back with solutions. It was even worse than normal Claude chat where the tool calls are collapsed. I have to assume this was done to show it "working", well I hated it.

After churning for a bit with very confusing UI [0] it's findings were not even "lackluster", just completely useless. I have annual performance reviews I need to review and it offered to help. I didn't expect anything useful and didn't plan on using anything it did (I do performance reviews by hand for obvious reasons) but I wanted to see how it'd do. Well, it didn't have access to the most important parts of the data but instead of asking or realizing it had incomplete data it tried to write reviews from scratch using my Obsidian notes. Good idea I guess? But it ignored the employee-filled-out review and the layout I need to use. Again, it didn't have access but instead of asking "is there a form/rubric/etc" it just forged ahead in an effort to show its usefulness which only highlighted its incompetence.

The other 2 tasks it tried to help with included a failed "get ready for product launch" document that had very little to do with what we are planning to launch, complete throwaway, and it suggested scheduling a meeting with an email spammer. "XXXX Says they're free to meet but you haven't set a time yet" - and I never will. PS: HN spamming has gotten a little out of control "I saw your comment on <LLM generated> and I'd love to talk to you about how <insert product I'm shilling for> can help!".

So yeah, a pretty terrible start to Duet. To be fair, Dots was a trainwreck when I tried to use it to write code. I don't love codex but Dots+coding was like the dumbest codex with less control over what it's doing (or more often, not doing). Really, I cannot stress enough how much Dots sucks at writing/overseeing-writing code. But the sweet spot for Dots was monitoring communications and raising things I might have missed, like emails/slack messages that slipped through the cracks or were filtered out but actually were something I needed to look at.

[0] https://cs.joshstrange.com/W2VqC23v It did this 6 more times... completely useless.

← PreviousPage 2 of 34Next →