AI has fundamentally changed the way we work, but the real value comes from working with other humans. This is why I created Wind Stream. AI for the way teams actually work.
https://t.co/kIHu994cFA
The problem with the "if it works who cares what the code looks like" mindset for agentic work is that it assumes the agent has a perfect understanding of "works." Realistically, things are underspecified, agents make bad assumptions, etc.
To be fair, agents are pretty good at unit test coverage. They're pretty bad at designing human experiences (API, CLI flags, etc.), especially cohesive ones for future roadmap plans they may not have visibility into (unless your backlog is perfect and vision fully laid out, which I doubt). They're bad at knowing where performance matters and what type (CPU vs memory tradeoffs). They're bad at where compatibility matters and where it doesn't (and tend to err on the side of preserving it without further guidance). Etc.
Unless you have this ALL specified, you can't possibly claim "it works" without taking a look and thinking about it.
I'm certain Fable 5 is more capable, but do I want to talk to it?
Opus 4.8 is downright neurotic. If I can get it a well defined task, it's amazing. But wow, it's not a talker.
VoidZero is joining Cloudflare.
Our mission stays the same: to make JavaScript developers more productive than ever before. Vite, Vitest, Rolldown, Oxc, and Vite+ remain MIT-licensed. Evan and the VoidZero team will continue leading them.
Cloudflare shares our commitment to open source. Together, we can keep investing in the tooling developers rely on every day, while bringing the Vite ecosystem and Cloudflareโs platform even closer together.
@zeeg Nope. Confirmed here. We reverted back to 4.6 with 4.7 for development. So far "I don't like" 4.8, but it's very productive so I haven't reverted.
The harness improvements we made to make 4.7 align with our platform charter have held through 4.8. Opus 4.5 is the gold standard.
@Jaytel Interestingly enough, 4.7 in a heavily tuned harness is inoffensive. The migration (I won't call it an upgrade) from 4.6 to 4.7 required considerable tuning. Made the harness much better though. That's the only benefit I've seen so far with 4.7.
GPT-image-2 benchmarks higher than Nano Banana 2, but my testing doesn't back up real output. It's an order of magnitude better in terms of quality, speed, and ease of use.
I've been building something that solves real pain we're all experiencing working across teams with AI today. I'd love to see if it aligns with your experience.
Ghostty is leaving GitHub. I'm GitHub user 1299, joined Feb 2008. I've visited GitHub almost every single day for over 18 years. It's never been a question for me where I'd put my projects: always GitHub. I'm super sad to say this, but its time to go. https://t.co/DQDemHdytV
I've been working on something really cool with @SaraD , which you'll hear a LOT MORE very soon. We'd love to hear your feedback on how AI fits into your current workflow.
Please retweet for reach. ๐ฑ
https://t.co/pncjUXGKBJ ๐ฟ