I did too, those awstrack.me URL's look super suspicious and I hadn't seen this alert trigger before so didn't know what to expect.
At first I was sure it was a phishing attempt. Then went to the console (not using those links)
Saw there was an outage where the console was wrong (no mention of email alerts)
Then I thought I was hacked - what a perfect cover up for someone to evade detection when the console was wrong.
Looked at some logs, realized the incident text was just not exhaustive on the impact.
Went back to my cup of coffee.
Sounds interesting, I wonder if any of it will be available on consumer hardware. Making CUDA work across the stack seems to be the key thing that helped NVIDIA become the main purveyor of GPUs for AI. Apple seems to be doing well by the same approach.
Last time I looked, anything remotely useful here was only on the Xeon chips, which means there'll be very few open source and even fewer non-corporate folks doing anything interesting with it.
AMD seems to be better at that, maybe there's hope.
this article sounds like a poorly AI written of a mashup of the https://www.acquired.fm/episodes/costco and CNN news article on the Mamdani bodega killing grocery store (which sounds super expensive for the value it provides)
and webvan back in 2000 - both amazon's attempt and webvan are unfortunately gone - it's cheaper to throw away packaging away, and that's super unfortunate and sad.
Congrats to the new outlook team on the performance improvements, I certain it used to take 30 seconds to do it, and they've cut it to 10s !!
But seriously, can we please make desktop productivity apps not suck on windows? I started programming on windows, old school Win32 with a little MFC. Still have the super thick MFC book from MikeB somewhere in the closet. It was better than the alternatives at the time.
Now I look at the windows developer site and I can't even figure out what happened since I stopped Win32 programming at around 2004. It's a total train wreck of abandoned technology, each worse than the previous ones.
Office (and to some degree visual studio), used to be the lighthouse, best in breed application, often using api's that were not yet public and styles that were not yet adopted. I remember buying component libraries that emulated these to make better looking and performing apps.
I'd look at windows again if they would make apps not suck and be ones that the industry strives to emulate. Without that, Linux or Mac is just as good (actually better since they have decent userlands).
I was speaking more on the linear agent vs the existing integrations. We love the linear guided reviews and issue tracking so have high hopes on getting a good DX here from y'all.
The claude integration - works as well as anthropic will let it work, since you can either automate it (anywhere from 'claude -p xxxx' or the api and ignore your subscription and pay by the token, or open a crapton of tabs with the terminal, or paste it into a bunch of sessions in their app. Which works more-or-less but it's cheaper than per-token costs.
The linear agent, doesn't seem to read the AGENTS.md file, follow along on a PR nor nor let you configure a sandbox (it told me this:
```
Note: I couldn't run ruff/pytest here (no uv/venv in the sandbox), so I verified syntax via AST parse only. The Postgres-backed tests will run in CI.
```
After I asked it to look at the PR check failures.
To be fair, claude code does it 70% of the time (the other 30% the sandbox is dead), and cursor about 10% of the time.
Yes, I still use it, although less than I would otherwise.
Good:
- Composer 2.5 is pretty decent for the quality / price ratio.
- Easy to assign an issue to it in Linear (I know Linear just added this natively for linear agent, but it seems rubbish compared to Cursor)
- Bugbot actually finds some useful issues (things Claude and Codex will miss)
- Using @cursor in github usually works well, and better than @copilot.
- Working with Python Monorepos with UV in their IDE. VSCode and Cursor work well here (Antigravity managed to screw it up somehow).
The Bad:
- Usage/billing dashboards - These are are opaque and you can't attribute what actions map to what spend.
- cursor won't follow PRs well like Claude Codes does.
- Setting up environments is less good than Claude Code
- Their IDE fork is woefully out of date, it'd be nice if it had more of the codeium fixes.
The Ugly:
- Settings - Try to turn off bugbot, there's multiple places you have to do it. Good luck figuring them all out.
- Support - they are polite, but gas light you and tell you it's your fault their product's settings are awful.
This happens with me every year with Barnes and Noble bookstore. I canceled their annual membership back in like 2018, Starting in 2021 this zombie account started charging me for a renewal. I reached out to support and they disavowed any knowledge of how this could happen, had no record of charging me, and no clue how it could have happened.
Since Amex is nice about allowing someone who charged you in the past - keep charging you even as your card has changed, they allow the charge.
Every year I do a chargeback, every year. Amex can't figure out how to not allow it and B&N has no idea why they charge it. Hmmm...
WoW. That's certainly a surprise to me. I'd never expect an invoice after not putting in a card.
I also believe this is totally just a case of "billing and metering is hard, and may actually be a larger engineering effort than your actual service".
I was just looking at them earlier today since our Github actions are slow AF, and while they sounds great, this tells me it'll cost me more time to make sure I babysit it than most other trials.
With most of these, they end, the service stops working, and you have a choice to make: (a) it was worth it sign up, (b) not worth it revert.
"""
Steve Ballmer
In IBM there's a religion in software that says you have to count K-LOCs, and a K-LOC is a thousand line of code. How big a project is it? Oh, it's sort of a 10K-LOC project. This is a 20K-LOCer. And this is 5OK-LOCs. And IBM wanted to sort of make it the religion about how we got paid. How much money we made off OS 2, how much they did. How many K-LOCs did you do? And we kept trying to convince them - hey, if we have - a developer's got a good idea and he can get something done in 4K-LOCs instead of 20K-LOCs, should we make less money? Because he's made something smaller and faster, less KLOC. K-LOCs, K-LOCs, that's the methodology. Ugh anyway, that always makes my back just crinkle up at the thought of the whole thing.
"""
I have to say the new model is quite good at the basics, I've been handing over more and more tasks from Linear straight to it instead of the copy-paste into Claude dance lately.
At this point, more of my complaints are on the harness side, which is odd since originally they were by far the best harness out there.
Support - This is pretty much non-existant, it's community support or sales support.
Interacting with GitHub - this should work and be awesome, Claude code does this well (responding to lint errors and comments). Cursor you have to poke the agent to look at the comments or lint errors, and even then it's about 10% good. Even GitHub Copilot is better here.
Bugbot - I have it setup to trigger manually, but it still seems to wake up and burn 80-120k tokens just to notice it's configured to be manually invoked. When it does run, it tells me there's no issues (but claude or copilot both find real things)
App - When you have both agent window and the ide windows, it's hard to open up the code in the right directory. A simple "cursor ." from the terminal used to do it, now it'll often open the agent window, you have to try a few times for it to work.
I love that they are running super fast, it's just hard when many of the basics break or don't work.