@maxedapps
Account based inGermany
About this account
- Account based in
- Germany
- Connected via
- Germany App Store
Account-level information from X, not a live location or the device used for a specific post.
I teach AI & coding: https://nitter.cf/t.co/Pnlz9drNJ7. NEW program: https://nitter.cf/t.co/jq7sTsHllp. And I do YouTube https://nitter.cf/t.co/i1tUc3y6mn
Joined August 2015
- Tweets7.7K
- Following190
- Followers69.1K
- Likes11.2K
Ultimately, we'll land at an email-like system for managing agents.
Message out, handle results in the inbox.
If you treat them like coworkers, you should work with them like you do with coworkers.
Lol
Antigravity 2.0 now features a dedicated planning mode, just like the Antigravity CLI.
Type /plan and the agent steps back to think through the task, and conduct extensive research, before generating an implementation plan for your review. The agent will ask for your approval before it begins executing on the task.
You can also ask for a plan naturally in your prompt, for a lighter version of a plan.
All nice, but they'll first have to ship a model on par with Opus 5.5 if they want to sell it
So the ChatGPT Pro max plan at $500 will not only include higher rates but also faster inference.
„Fastest work in codex“
Opus 5.5 is truly a model from another world.
I can’t express how much this model blows me away.
For me, it feels like a bigger step than even Fable felt back then.
Unbelievable
This must be one of the saddest charts I’ve seen @patrickc
I couldn’t agree more.
And it’s very true: we’re essentially the „early adopters“. It’s crazy if you talk to „normal“ people and they have no idea what’s coming for them
I've been reflecting on this. I'm generally a very optimistic person, but lately I've been feeling a strange sense of sadness that coding has been "solved".
I can, and have, made many caveats to "solved"... but I also know that in five years none of them will matter. It will be solved.
Building things is still fun, but a different kind of fun. It doesn't activate the same parts of my brain as trad coding did.
I think it's okay to mourn the thing we loved doing being different now. Software was the first job to go through the five AI stages of grief, which means we're also some of the first to internalize that this will happen to many other jobs.
It was the best of times, it was the worst of times.
Maximilian retweeted
I've been reflecting on this. I'm generally a very optimistic person, but lately I've been feeling a strange sense of sadness that coding has been "solved".
I can, and have, made many caveats to "solved"... but I also know that in five years none of them will matter. It will be solved.
Building things is still fun, but a different kind of fun. It doesn't activate the same parts of my brain as trad coding did.
I think it's okay to mourn the thing we loved doing being different now. Software was the first job to go through the five AI stages of grief, which means we're also some of the first to internalize that this will happen to many other jobs.
It was the best of times, it was the worst of times.
Got Opus 5.5, GPT 6 Sol, Grok 4.7 over those last ~30 hours. AND also MiMo 2.6 which is maybe not the same league but also pretty impressive!
Long story short: You should use Opus 5.5. If you don't have a Claude sub, now may be the time to get one.
If you're stuck with the OpenAI sub, you're out of Claude usage, or you simply prefer OpenAI, you should still go with Astra, even though it's a bit harder to "wield" and get useful results out of.
GPT 6 Sol is good if you need to go farther (usage wise) than with Astra. But it IS simply a smarter 5.6 Sol. Which, I guess, was the goal.
It just seems bad because Opus 5.5 launched yesterday, too.
By the way, don't sleep on Luna! It's dirt cheap, even with max effort. Obviously not the model you should use for frontier coding work, but for very trivial tasks, or especially for non-coding tasks it may very well be the best choice. Because it's essentially infinite usage you get out of it.
Grok 4.7 is rather disappointing. Slow, inefficient, and doesn't feel that much smarter than 4.6. I'll still to 4.6 for now therefore. Well, actually I'll use Opus 5.5, but if I'm working with Grok, it'll be 4.6.
Hopefully 4.8 will be a step in the right direction then.
MiMo 2.6 is a Chinese (open weights) model that comes in "Pro" and "Flash" variants. I used the "Pro" one yesterday and found it really good.
To be honest: Not something I'll keep on using because I got my subscriptions with Claude etc.
BUT MiMo is incredibly cheap (I used it at API prices and only spent a few bucks) and pretty intelligent. The best open model right now. So give it a try if you got use-cases where you can't or don't want to use a subscription.
Be aware though, that MiMo 2.6 Pro at max effort is VERY slow and thinks A LOT.
Is the Whoop 5 good? For workout tracking, sleep tracking, daily activity tracking? Can I get access to my data via API and build my own dashboards etc?
Website sounds like all that is the case but I want to hear from people that actually used it
Every now and then you get a model that looks like a real game changer again.
Opus 5.5 really seems to be such a model.
Still in early testing (I did not get any early access) but it's precise, quick, affordable, it writes well (!!!) and so much more.
Anthropic really cooked here.
