Seems to be the case that many users find Claude Code to be the most capable but less trustworthy than its competitors. We're hopeful that tools like cc-canary help users gain that trust back so they can continue to use the best tools on the market.
Wow, bad enough for them to actually publish something and not cryptic tweets from employees.
Damage is done for me though. Even just one of these things (messing with adaptive thinking) is enough for me to not trust them anymore. And then their A/B testing this week on pricing.
The A/B testing is by far the most objectionable thing from them so far in my opinion, if only because of how terrible it would be for something like that to be standard for subscriptions. I'd argue that it's not even A/B testing of pricing but silently giving a subset of users an entirely different product than they signed up for; it would be like if 2% of Netflix customers had full-screen ads pop up and cover the videos randomly throughout a show. Historically the only thing stopping companies from extraordinarily user-hostile decisions has been public outcry, but limiting it to a small subset of users seems like it's intentionally designed to try to limit the PR consequences.
The best possible situation that I can imagine is that Anthropic just wanted to measure how much value does Claude Code have for Pro users and didn't mean to change the plan itself (so those users would get CC as a "bonus"), but that alone is already questionable to start with.
People come at this with all kinds of life experience. The above notion of trust to me is quaint and simplistic. I suggest another way to frame trust as a more open ended question:
To what degree do I predict another person/org will give me what I need and why?
This shifts "trust" away from all or nothing and it gets me thinking about things like "what are the moving parts?" and "what are the incentives" and "what is my plan B?".
In my life experience, looking back, when I've found myself swinging from "high trust" to "low trust" the change was usually rooted in my expectations; it was usually rooted in me having a naive understanding of the world that was rudely shattered.
Will you force trust to be a bit? Or can you admit a probability distribution? Bits (true/false or yes/no or trust/don't trust) thrash wildly. Bayesians update incrementally: this is (a) more pleasant; (b) more correct; (c) more curious; (d) easier to compare notes with others.
I "subconsciously" moved to codex back in mid Feb from CC and it's been so freaking awesome. I don't think it's as good at UI, but man is it thorough and able to gather the right context to find solutions.
I use "subconsciously" in quotes because I don't remember exactly why I did it, but it aligns with the degradation of their service so it feels like that probably has something to do with it even though I didn't realize it at the time.
Anthropic definitely takes the cake when it comes to UI related activities (pulling in and properly applying Figma elements, understanding UI related prompts and properly executing on it, etc), and I say this as a designer with a personal Codex subscription.
it's been frustrating how bad it is at UI. I'm starting to test out using their image2 for UI and then handing it to codex to build out the images into code and I'm impressed and relieved so far
Codex does better if you ask it to take screenshots and critique its own UI work and iterate. It rarely one-shots something I like but it can get there in steps.
I went with MiniMax. The token plans are over what I currently need, 4500 messages per 5h, 45000 messages per week for 40$. I can run multiple agents and they don't think for 5-10 minutes like Sonnet did. Also I can finally see the thinking process while Anthropic chose to hide it all from me.
Gemini fairs poorly at tool use, even in its own CLI and even in Antigravity. It gets into a mess just editing source files, it's tragic because it's actually not a bad model otherwise.
It frequently fails to apply its diffs at first but it always succeeds eventually for me. I'm happy with it. I understand it is slower than other models but it also costs barely anything per month.
This is the way if we can ensure manufacturing of the parts. It won’t catch on but it would be awesome to have “base” tractors that are mechanical and predictable. Then you slap on whatever software on top that helps (automation, etc). But they need to be decoupled imo.
i have a farmall hand cranked tractor, going on 90 years old, so far its been rubber parts, and clutch pads.
as far as auto mation goes, thats how implements used to work. it was a tracter/thresher/combine. then a bale counter is slapped on then maybe row sighting or guidance, etc.
if your really snazzy, the implement is actually mapping the soil for moisture, or rough composistion and holding data to use in reformulating or notating your current cultural plans, i.e. supplemental spot feeding and irrigation.
I still got a farmall 230, super easy to fix and maintain and works perfect for my small bit of land. An electric starter addon is really nice for winter starts though instead of killing your arm.
While I’m not at all surprised that they’re still running, I am a little surprised at how many Farm-all owners are on HN. Farm-all H owner checking in :)
And how many acres are you farming on it? Today's world of agriculture is much higher tech-based (for many good reasons, primarily yield) than back in the horse and buggy days of farming.
I know of a forklift that's pushing 80 and still used in a lumber yard (i.e. a material handling centric workplace)
Other than ~30min it takes to teach an employee to drive manual it doesn't do anything worse than the modern ones it works alongside and it does a handful of minor things much better by virtue of predating OSHA.
Wasn't designed with the assumption that when someone gets hurt lawsuits and fines could/would fly and so there's a lot less of the manufacturer covering it's ass in the design.
There's nothing to keep you from putting it on two wheels and the transition from on the ground to off the ground is pretty graceful and operator friendly whereas the newer lifts rely on a pressure release valve to keep them from lifting that much and (presumably) because they were always expected to be far from ragged edge their weight distribution is not really proper for that. The counterweight is substantially taller so how hard it pushes down is reduces more quickly as the lift comes up so it lifts tire further (and is more likely to dump the load or go over). This also means the old lift has a way lower ass pucker factor when doing stuff at max height. The real nuisance is when braking though. Yeah you "shouldn't" brake with the load up but operators who get good will raise the load at speed as they are coming in to put a pallet of stuff on top of another pallet of stuff and then when they brake it can get sketchy. The new lifts do corner much better unloaded though so I guess you could be much faster zipping through a warehouse on a new lift (but what workplace would permit that? And top speeds are about the same so there's no benefit in a big outdoor workplace like say an airport or shipyard).
There's no seat switch or other safety interlocks so you aren't putting a ton of wear on it if you're constantly getting out to fiddle with stuff. This also means you can do "unsafe" things like stand beside it and wrangle something and just reach in and make the mast go up and down. While in a textbook world this is "bad" and you "should only pick pallets" and "everything should be strapped to the pallet" in the real world you make all that back and more because it means you can use the forklift as a glorified engine hoist/shop crane without a helper. Hook and chain operations are made much safer/more reliable by this too since the operator can be sure things are good and is not tempted to half ass it to save the time of getting back out. Sure you could always add a helper but that's dangerous too because one person doing stuff near equipment and one person running equipment opens the door to miscommunication related injury that can't really happen among one person.
I'm sure "at scale" the new lift is safer, but safer for who? In what operating context? How big is the difference?
This is what a "bobcat" has become for UGV startups. It's a low tech proven platform that you can basically modify with attachments to do a lot of UGV work.
I was assuming the same. This might be fine for a small setup but I'd imagine all the digitization shenanigans was done so efficiency could increase. I imagine for large scale operations this would be like replacing your steam engine with a horse.
> “When we work on making our devices accessible by the blind, I don’t consider the bloody ROI.”
I just have to call out how much this impacted my mom’s life. She’s 100% blind and has access because of her iPhone and iPad. Yes she learned JAWSs and literally took classes to do it. Every single windows update has made it so she’d have to retake this class. The iOS updates a rocky but she isn’t literally hamstrung.
My dad, damn near 80, is still happily using his 2012 i7 Mac mini I set him up with before moving away.
Blind person using Apple products here, and at least for phones, I agree. I wouldn't say it's exclusively because of iPhone, but a large part of my independence is definitely it. There have been problems, bugs that go unfixed for years, MacOS VoiceOver is quite a disaster even though I do still use and enjoy the platform overall, and anything worth using can be criticized I think. But iOS has so many features built in that help me every single day. VoiceOver, but also all of the features utilizing vision like door detection, OCR, etc. they're in the magnifier as well so you don't need VoiceOver enabled to play with them, and I think a number of them also require a lidar sensor?
Anyway, my phone is such an important companion wherever I go that I keep several magsafe batteries on me whenever I leave the house for a significant time. It has made an absolutely huge difference in confidence. It is definitely one of the single most important assistive tech devices I have together with my computer.
It is just random bugs. Switching punctuation schemes. The terminal doesn't read very well, VoiceOver loves to say "not responding" in Safari and locks up, live regions don't always read correctly, quick nav (basically automatically holding down the voiceover modifier so you can more quickly use navigate through the screen) adds random delay to each key press, it's just lots and lots and lots of small issues like this that compound. This is just a small list of them. None of them are a huge problem by itself, but combined they do make things frustrating sometimes. And then of course the ability to script badly behaving, or completely inaccessible, apps is just missing, so you can't fix apps even if you knew how to.
And of course VoiceOver on the Mac is all you get. So if you don't like it, tough luck. You won't ever get a real alternative that can access what VoiceOver can.
I honestly don't know how anyone thinks terminal support in VoiceOver is acceptable--it's virtually unusable. It's so bad that I used to fire up a Windows VM just for a functioning terminal, and while I was at it, I'd browse the web and use Notepad++ there, because Windows accessibility is just better (I used NVDA). But then I discovered Fenrir, figured out that it worked with Vim (NVDA doesn't), shut down the VM, and never looked back. Today, I use Wezterm, which VoiceOver doesn't read at all. In my case, that's good, because the only thing I want talking in the terminal is my terminal screen reader (I started writing my own, and it's my daily driver).
To be fair, reading the terminal is a completely different beast from reading a GUI. In addition to building a static view of the screen for review, you have to handle dynamic updates (auto read). Cursor movement tracking, figuring out when to read what, when not to read (an f just appeared on my screen, but I just typed the letter f; if key echo is turned off, I don't want to hear "f"). If a line was just added, it should be read, but if my cursor was moved to a different line, I want to hear the line it moved to, but not if that line was just read because it just appeared. All sorts of rules you sort of discover as you go. But the one thing you definitely don't want is for any new change to interrupt what was already being read, and that's exactly what VoiceOver does.
> VoiceOver loves to say "not responding" in Safari and locks up,
I wonder, what's the correct solution for this ? Because so many apps I use including browser are definitely "not responding" multiple times per day for various reasons (full ram, internet stall, etc.)
Using VoiceOver compounds the not responding issue. I don't know how its internals work, but I imagine it tries to keep a view of the window's state--tree of elements, ETC. If the window has a lot going on, VoiceOver can get really sluggish, and I think it must somehow block the underlying app's ability to send/receive events, because you will press VO+right arrow to move to the next element, VO says "Safari/Chrome/Brave" not responding, and if you open up Force Quit, it reflects the same there. Reading a large diff on GitHub flat out doesn't work for me at all. Also, sometimes when navigating certain webpages, VoiceOver will just outright crash. Luckily, it does restart itself (not that pressing CMD+F5 is hard), but then my focus is moved to a completely different part of the page.
Will drop this here in case you’re not aware of it (but I’m guessing you probably are), sorry if a bit off-topic.
I’m low-vision and made great use of Microsoft Soundscape until it got discontinued. I’d been waiting for an alternative for ages and didn’t realise one actually got released and is on the app store!
I absolutely LOVE! Voice Vista. It is an amazing bit of software. I wasn't able to use SoundScape when it first came out because it was never made available in my region, but VV is, and I would never want to miss it anymore when traveling. I love it. A lot.
For what it's worth, text selection has been badly broken on iOS for at least a decade and autocorrect has been steadily getting worse for probably the same amount of time, and these are features that affect the mainstream segment of Apple users on a daily basis. Apple seems generally happy to let bugs go unaddressed for years and years regardless of how many people they affect or how often.
It’s really really inconsistent. Sometimes select all is available, sometimes not. Sometimes the handles don’t work. Selecting text in a scrollable region is fiddly, etc.
I’ve seen an insane drop in the quality of swipe typing recently as well. To the point where I’ll often go back to regular typing. I’ve made maybe six or more corrections just to this paragraph alone.
I think swipe typing suggests words inconsistent with any higher level language model, like word tuples, when proposing words which are possible matches for letter sequences swiped.
and it drives me crazy too.
I've just had good luck it seems with text select.
Have you found any way to do a Find within a span of text on iOS? That would be very useful, but I haven't seen it.
Excuse my language here but: I fucking love this! My mom pretty much mirrors your experience. I purposefully left out macOS and voiceover. I would almost call it unusable, sadly. The amount of key layering that voiceover and macOS in general has makes it very hard to use.
I’ve been hacking on a macOS app that leans on LLMs, vision use, and the AX macOS APIs to try and make voiceover less.. prickly haha. Hoping to visit in person soon to watch her use it :)
The first time I saw a blind person using an iPhone, I was blown away. I follow some Apple engineers who work on accessibility, and they all seem very passionate about their work. It’s an area where I truly believe Apple is doing it to help people, not just for profit.
Just an aside, it's a ton of work to make accessibility work on anything other than the most native looking apps, as different settings will move the UI unexpectedly and creates a lot of issues to be taken care of because of the screen size and different layouts.
The screen speak for example, sometimes you have to manually make sure they speak in the right order because of UI elements are placed non-standard way like if you have a label as name, and one as phone number side by side, the speak may start going down vertically, and you have to fix it by grouping it or force it speak it manually. Small example.
I’m not blind, but I’m using accessibility features like Speak Screen, and the text-to-speech is pretty poor (mispronunciations abound, markup is ignored, punctuation is misinterpreted), usability is poor (can’t start at a user-selected location on a page for example), and it’s rather buggy, especially within Safari. It’s been that way for years, and it doesn’t seem like anyone at Apple is interested in making it a better experience.
When you have Apple level margins then you can definitely consider long term ROI (such as this entire thread, for example). Long term greedy, as they say.
I struggle not to have a cynical take these days. Of course he cared about the ROI. The ROI is access to an underserved market, a halo effect, a new community of adherents, a new reason for customers to cross the moat into the ecosystem… a modest investment with a durable long term return in multiple categories.
I appreciate that it’s a win-win for Apple and for its customers, and I firmly believe that accessibility features serve everyone eventually. I’m glad that there are some billionaires who also see it that way.
I guess I just wish we didn’t have to rely on rare cases of billionaires finding it in their own best interest to happen to serve the rest of us. Especially when the actual accessibility work and everything else is actually done by a whole class of people that never make headlines just for leaving their jobs and being replaced.
You're arguing that the action had some positive effects and therefore it was ROI positive. That doesn't remotely follow.
And most companies did NOT make the choice to be as accessible as Apple, which rebuts your theory that this was done only for the ROI.
Effectively you're so cynical that there's nothing Tim Cook could say or do that would convince you he was ever acting sincerely. It is comfortable to blame and rage but it is hardly good analysis.
It's obvious he has to be somewhat concerned about the ROI (or LOI) - if it cost ten times the value of the company to implement accessibility for the blind, it's not going to get implemented.
But the whole point of leadership should be to say "this doesn't bean count out perfectly, but we'll do it".
I get what you’re saying but in my 15 year career the ONLY time I was allowed to meaningfully work on accessibility was when visa hired me to remediate visa checkout. And that was literally because a tier 1 bank was going to drop their contract over it.
The ROI Apple will get is when all of us turn 70 and need these features we’re ignoring now
He told the shareholder entity to get out of AAPL stock if they cared so much about ROI. That doesn't sound insincere.
I'm looking for a recording of that shareholder meeting, to see if he looked and sounded insincere at that time but YouTube is insisting on showing the latest news.
It’s funny, a few months ago I would have been pretty excited about this. But I honestly don’t really care because I can’t trust Anthropic to not play games with this over the next month post release.
I just flat out don’t trust them. They’ve shown more than enough that they change things without telling users.
It might not be Cloudflare's fault, but it is their problem. If their customers can't use their products sporadically, it doesn't really matter why. Cloudflare taking a principled approach hurts their users in the short term, so they have to make a business trade-off between pragmatism and principle. Currently they're choosing principle, so it's reasonable to be angry at them for the short term issues that causes.
I didn’t comment on this when I saw it on threads/twitter. But it made it to HN, surprisingly.
I have a feeling these same people will complain “my model is so dumb!”. There’s a reason why Claude had that “you’re absolutely right!” for a while. Or codex’s “you’re right to push on this”.
We’re basically just gaslighting GPUs. That wall of text is kinda needed right now.
If you feel the need to do this, it’s time to move onto a tool you trust?
reply