~20 years working in tech for me, mostly big companies, and I’ve never been more miserable. I also can’t stop myself from using Claude Code and the like.
I think it’s a bit like a gambling addiction. I’m riding high the few times it pays off, but most of the time it feels like it’s just on the edge of paying off (working) and surely the next prompt will push it over the edge.
It feels like this to me too, whenever I give it a try. It's like a button you can push to spend 20 minutes and have a 50/50 chance of either solving the problem with effortless magic, or painfully wasting your time and learning nothing. But it feels like we all need to try and use it anyway just in case we're going to be obsolete without it somehow.
If it doesn't work, it's an annoyance and you have to argue with it.
If it does work, it's one more case where maybe with the right MCP plumbing and/or a slightly better model you might not be needed as part of this process.
Feels a bit lose-lose.
> I also can’t stop myself from using Claude Code and the like.
just.. uninstall it? i've removed all ai tooling from both personal+work devices and highly recommend it. there's no temptation to 'quickly pull up $app just to see' if it doesn't exist
It’s become a core expectation at work now (Meta). If I’m not actively using it, then I’ll be significantly dinged in performance reviews. Honestly, it’s forced me to consider going to work elsewhere.
It does _feel_ like the value and happiness will come some versions down the road when I can actually focus on orchestration, and not just bang my head on the table. That’s the main thing that keeps me from just removing it all in personal projects.
Is this coming from the hypothesis / prior that coding agents are a net negative and those who use them really are akin to gambling addicts that are just fooling themselves?
The OP is right and I feel this a lot: when Claude pulls me into a rabbit hole, convinces me it knows where to go, and then just constantly falls flat on its face and we waste like several hours together, with a lot of all caps prompts from me towards the end. These sessions last in a way that he mentions: "maybe its just a prompt away from working"
But I would never delete CC because there are plenty of other instances where it works excellent and accelerates things quite a lot. And additionally, I know we see a lot of "coding agents are getting worse!" and "METR study proves all you AI sycophants are deluding yourselves!" and I again understand where these come from, agree with some of the points they raise, but honestly: my own personal perception (which I argue is pretty well backed up by benchmarks and by Claude's own product data which we don't see -- I doubt they would roll out a launch without at least one or more A/B tests) is that coding agents are getting much better, and that as a verifiable domain these "we're running out of data!" problems just aren't relevant here. The same way alphago gets superhuman, so will these coding agents, it's just a matter of when, and I use them today because they are already useful to me.
no, this is coming from the fact OP states they are miserable. that is unsustainable. at the end of the day the more productive setup is the one that keeps you happy and in your chair long term, as you'll produce nothing if you are burnt out.
This is definitely the feeling i get. Sometimes it works amazingly well that I think "Oh may be the hype was right all along, have I become the old guy yelling at claude?" but the other times it fails spectacularly, adds a really nasty bug which everyone misses for a month or cant even find the file I find by searching.
I am also now experimenting with my own version of opencode and I change models a lot, and it helps me learn how each model fails at different tasks, and it also helps me figure out the most cost effective model for each task. I may have spent too much time on this.
I see not being able to install invasive kernel level anti-cheat as a positive. I uninstalled all Riot games before they rolled it out. I would’ve been pretty miffed if I had accidentally gotten their kernel modules simply because I wasn’t reading tech news before the auto update.
There are some clear objective signals that aren’t just user preference. I shelled out the $250 for Gemini’s top tier and am profoundly disappointed. I had forgotten that loops were still a thing. I’ve hit this multiple times in Gemini CLI, and in different projects. It gets stuck in a loop (as in the exact same, usually nonsense, message over and over) and the automated loop detection stops the whole operation. It also stops in the middle of an operation very frequently. I don’t hit either of these in Claude Code or Codex.
There certainly is some user preference, but the deal breakers are flat out shortcomings that other tools solved (in AI terms) long ago. I haven’t dealt with agent loops since March with any other tool.
That’s a pretty uncharitable take. Given the scale of their recent launches and amount of compute to make them work, it seems incredibly smooth. Edge cases always arise, and all the company/teams can really do is be responsive - which is exactly why I see happening.
A company with a literal embedded payment processor, including subscription services for half of all mobile users can't manage to take payments for their own public facing services seems like a huge fucking failure to me.
Especially for software developer and tech influencer focused markets.
Considering the product itself seems to be excessively limited without actually getting paid for it, and the paid tier itself having so many onboarding issues, as a critical usage path, it's pretty bad.
This is in a $3.6 Trillion company, for a product they're spending billions a quarter to develop, with specialized employees making mid 6-figure to 7-figure salaries and bonuses... you'd think somebody has the right connections into the departments that typically handle the payment systems.
My expectations shoot up dramatically for organizations that have all the funding they need to create something "insanely great" in terms of user experience the further they fall short... I don't know who the head of this group/project/department/product is... but someone failed at their job, and got payed excessively for this poor execution.
> It is mostly about shifting profits from mom and pop, low regulation hemp industry to wealthy corporations that own dispensaries that have gargantuan regulatory costs that gatekeep out most the competition.
That’s a big assertion that needs evidence. I’m strongly in favor of legalization but not deregulation. It was a pretty big loophole that allowed what’s essentially weed to sidestep the regulation their competitors faced - and there wasn’t great consumer awareness about the differences even though there were safety implications: https://drexel.edu/cannabis-research/research/research-highl...
This law seems pretty well targeted in its scope, bringing the 2018 law back to what was intended (easy legal CBD/hemp, as long as there aren’t other things in there).
There was absolutely no federal regulatory framework for marijuana. none. It's just plain illegal. Unless you can get one of a handful of research licenses, which is almost totally irrelevant.
Hemp had some, fairly weak regulation. And theoretically, testing requirements, although they were deferred and deferred to the point they were basically done only privately with the idea the DEA would eventually get involved.
Instead they're just dumped now into the marijuana bucket which has no federal regulation at all, or alternatively, at the state level the states could always define their regulatory framework to be agnostic to THC content of cannabis.
So this does the exact opposite of what you had hoped.
Yet Kratom is legal, yt is recommending to me some product called "meth" (not joking) and there are a million new research drugs coming out every decade?
It's just old-school think of the kids and not in my territory. We don't know how to regulate and handle this because our politicians and more and more our citizens don't understand what is being voted on or has been happening in their own states for 7 years.
Have you used these products? It's a shame, the quality that I was getting just within the past 3 months was incredible and it is market not afraid to try new stuff.
I'm sad, flower from OR, NC, OK, IN, and others will never legally hit my lungs. Back to the cartels? Or perhaps I should overpay by $200 with the comfort of having 0 clue where it comes from, again?
>(easy legal CBD/hemp, as long as there aren’t other things in there)
Your ignorance shows in spades. The arbitrary ban on THC and its analogues prevent chronic pain patients like me (a criminally underserved market) from becoming addicted to the big pharma system. The "other things in there" argument is the same as razorblades in candy, sanctimony to portray dissent as degeneracy.
In my experience it is also the other things in there which helps with the pain relief. Doctors in my country talk about the entourage effect and mixing strains as they reckon it's not just the THC which is helping.
I can imagine people in the future looking at us like idiots as they use cannabinoids in the same way we use paracetamol.
From personal experience suffering from chronic pain cannabis is absolutely transformative. The difference between a life spiralling to nothing just about surviving on opioids compared to effective pain relief from cannabis and being able to work and be productive again.
One of the tragedies of the 20th and hopefully not the 21st century. So many people in so much unnecessary pain.
Looking at history I could quite easily come to the conclusion... ...due to racism.
> From personal experience suffering from chronic pain cannabis is absolutely transformative.
Like all drugs, it’s sad it doesn’t work this way for everyone. I had to transition from cannabis to opiates and lyrica. I wish this was not the case.
They suspect it’s due to the source of the pain (spinal cord injury) and the cannabis is “exciting” my nerves in the wrong way, as it actually increases my pain; or at least my perception of it.
Selling it as a pain reliever I can't buy into personally based off my anecdotal experience. I've had chronic pericarditis for more than a decade now and THC amplifies mine as well, as I tend to focus more on the pain. I think it's a very subjective thing, depending on many factors; strain, type of pain, person, etc.
AT&T is pretty bad in its own way. They snoop DNS and to sell your info (including physical address) to advertisers - even if you switch your DNS providers. They used to had a paid opt out (~$20/mo IIRC) but I don’t see that option anymore.
This is quite easy to avoid by using DNS over TLS. It's like 15 minutes of effort in some OpenWRT documentation [1]. If you want any hope of having some semblance of control and privacy, you would already be using your own router, with their CPE being relegated to modem-only duties. It only makes sense that in this situation you choose a router that can run highly-configurable and privacy-preserving software.
I did it several months ago, including the optional adding an outbound firewall rule dropping forwarded UDP/TCP 53 traffic (I tried the redirect rule suggested there first, but it didn't work and the firewall ruleset failed to load, so a drop will have to do. I didn't bother investigating why, because everything on my LANs is configured to use the router as their only nameserver anyway).
I also added a rule dropping it from the router itself in case something breaks, for example if it suddenly decides to start honouring the DHCP-received nameserver addresses (my ISP) despite being configured not to.
EDIT: The article doesn't make this clear, but the bootstrap section is only necessary if you specify upstream nameservers by name (e.g. "https://dns.cloudflare.com/dns-query"). This is not required. For example, you can configure a manual upstream of "tls://1.1.1.1" like I did, and then it doesn't need to do any DNS lookups at all, so does not need to be configured with bootstrap servers, so will not break if you add the 2 firewall rules I mentioned.
I wasn't really meaning to defend AT&T as a good option, just a slightly less evil one. I'm surprised I have a choice at all out here in the sticks. A lot of places just have one provider.
This reply seems to miss the question, or at least doesn’t answer it clearly. Is this service overly tolerant of mispronunciations? Foundational models are becoming more tolerant, not less, over time which is the opposite of what I’d want in this case.
It's less tolerant of mispronunciations. There is custom promting to explicitly leave in mistakes and to not fix them. It's still not perfect and it (the speech to text module) sometimes corrects the user's pronunciation mistakes.
I’ve found such changes can actually be a draw at first. “Hey look, progress and activity!”. Doubly so as a primarily C++ dev frustrated with legacy choices in stl. But as you and others point out, living with these changes is a huge pain.
It’s definitely not a transaction. Every time I’ve seen push come to shove, companies prioritize the folks they see as critical to their company’s success with loyalty not even being a small factor. And if it’s a moderate to large sized company, many of the decisions will be made by a consulting firm with 0 context (or care) for loyalty.
Haven’t read the paper but my guess around that is that the same reason sparse attention networks (where they 0 out many weights) just have the sparse tensors be larger.
I think it’s a bit like a gambling addiction. I’m riding high the few times it pays off, but most of the time it feels like it’s just on the edge of paying off (working) and surely the next prompt will push it over the edge.
reply