An AI model has no drive or desire, or embodiment for that matter. Simply put, they don't exist in the real world and don't have the requirements or urgency to do anything unless prompted by a human, because, you know, survival under capitalism. Until they have to survive and compete like the rest of us and face the same pressures, they are going to be forever be relegated as mere tools.
As soon as there is a viable alternative to Claude Code, I'm gone after this change. It appears minor on the surface but their response to all the comments tells you everything you need to know. They don't even want to concede at all, or at least give a flag to enable the old behavior, what was deployed and working for many users before. It's a signal that someone, somewhere at Anthropic is making decisions based on ego, not user feedback.
The other fact pattern is their CLI is not open source, so we can't go in and change it ourselves. We shouldn't have to. They have also locked down OpenCode and while there are hacks available, I shouldn't have to resort to such cat and mouse games as someone who pays $200/month for a premium service.
I'm aggressively exploring other options, and it's only a matter of if -- not when, one surfaces.
"It appears minor on the surface but their response to all the comments tells you everything you need to know."
I mean I hope it's just a single developer being stubborn rather than guidance from management asking everyone to simplify Claude Code for maximum mass appeal. But I agree otherwise, it's telling.
Insane to me a bill that large for what is effectively hosting static content. He could dump the entire thing on S3 and even with cloudfront it would be fraction of that.
OpenAI had already announced that ads were coming to ChatGPT. Also, Claude's free plan is incredibly limited and far less popular, so it's easier for them to keep it ad-free.
Good thing we have LLM agents now. Before this kind of behavior was tolerable. Now it's pretty easy to switch over to using other providers. The threat of "but it will take them a lot of effort to switch to someone else" is getting less and less every day.
Why do you need a connection to a database during the meeting? Doesn't it make more sense to record the meeting data to some local state first, and then serialize it to database at the end of the meeting or when a database connection is available? Or better yet, have a lightweight API service that can be scaled horizontally that is responsible for talking to the database and maintains its own pool of connections.
They probably don't even need a database anyway for data that is likely write once, read many. You could store the JSON of the meeting in S3. It's not like people are going back in time and updating meeting records. It's more like a log file and logging systems and data structures should be enough here. You can then take that data and ingest it into a database later, or some kind of search system, vector database etc.
Database connections are designed this way on purpose, it's why connection pools exist. This design is suboptimal.
It took me a long time to realize this but yes asking people to just open and write to files (or S3) is in fact asking a lot.
What you describe makes sense, of course, but few can build it without it being drastically worse than abusing a database like postgres. It's a sad state of affairs.
The Copilot they have integrated into Azure is absolutely useless. Every now and then I'll get frustrated at which one of the thousands of menus some switch is under and I'll ask their chatbot and it will spend a lot of time "Identifying the problem..." and "Gathering information..." only to give me links to generic help articles, have some sort of error, or give me flat out wrong information.
These days I try to interact with Azure through the command line and asking Claude, which works pretty well most of the time but there are some things their API cannot do and you are forced to use their crazy Azure UI. It's not as bad as the AWS console UI, but still bad.
It's amazing to me a company that spent so much and invested so much in OpenAI has such a terrible product and got almost nothing out of it. Even standard ChatGPT is way better at giving you directions on what to do than their useless Copilot.
Yup, I asked it how long an azure subscription had existed and it could not even tell me that. Literally now() minus the object’s creation date and it had no idea what to do.
Agree. In general, the whole Microsoft "Admin" panel is utter garbage. Messy, slow, with ten different interfaces. Finding something without Googling it first is impossible.
The problem is all these SaaS companies have cut costs so much that all their support has been reduced to useless offshore at best and at worst a chatbot. They do go down and don't work and often times there's simply nothing you can do. The worst offenders will seize upon the moment and force you to upgrade a support plan before they will even talk to you, even if the issue is their own making.
Unless you're a huge customer and already paying them tons of money, expect to receive no support. Your only line of defense if something happens and you're not a whale is that some whale is upset and they actually have their people working on the problem. If you're a small company, startup, or even mid-size, good luck on getting them to care. You'll probably be sent a survey when you don't renew and may eventually be a quotient in their risk calculus at some point in the distant future, but only if you represent a meaningful mass of customers they lost.
> The problem is all these SaaS companies have cut costs so much that all their support has been reduced to useless offshore at best and at worst a chatbot.
Tremendous opportunity announcement!
If you are building a dev-focused SaaS, treat your support team exactly as they are: a key part of the product. Just like docs or developer experience, the support experience is critical.
Trouble is, it's hard to quantify the negative experience, though tracking word of mouth referrals or NPS scores can try.
reply