@angular2react

Angular & React Team Lead by day, HTMX dev by night

angular2react.substack.com
Joined September 2022
  • Tweets1.1K
  • Following210
  • Followers34
244Media
Replying to @JackEllis
I stop, when AI stops gaslighting me! xitter.cf/angular2react/status/2…
angular2react@angular2reactSep 29
wtf
4
Replying to @manuelmaly
Works well for you? I'll test it out
1
19
Replying to @manuelmaly
Opus 5.5 -> Planning Haiku 5.5 -> Implementing Feels SOOO much better. Only downside, I have to use claude code and can't rely on Pi :-/
1
1
96
Replying to @manuelmaly
Not coding.
1
29
Replying to @BackersGamesF
I mean - good? Hopefully they one-up each other until it's not horrible slow? 5sec for a app to start?
212
Replying to @ryan_doser13
Similar vibe. news.ycombinator.com/item?id… It's OK if you don't understand the appeal. It was the same for me with Vercel. "It's a stupid AWS wrapper!?" - nah, that's not the interesting part people are willing to pay for.
9
Replying to @PaulSolt @pvncher @dkundel
At the moment, both are quite bad in comparison to Anthropic. xitter.cf/angular2react/status/2…
angular2react@angular2reactOct 13
Replying to @jedfrankowski @michael_kove @mattepstein
Open AI models are really bad atm. I really love my Pi agent, but it's so bad that I went back to claude. Results are much better. xitter.cf/angular2react/status/2…
21
Replying to @michael_kove @PositivFuturist
Obviously smart-fridges youtube.com/watch?v=BnKpNVHw…

Gilfoyle Saved Pied Piper - Silicon Valley

Gilfoyle has put Pied Piper library into Jian Yang's Smart Fridge a...

youtube.com
1
2
Replying to @jpschroeder
Having Astra so many points ahead of DeepSeek v4.1?
31
No lie detected. xitter.cf/TheGingerBill/status/2…
gingerBill@TheGingerBillOct 12
I wanted to show how janky the AI slop is of the FilmCraft software. I wanted to set the stage for how bad it is, and how little care there has actually been in implementing this tool, and most definitely the entire Craft suite too. It's all AI slop that does not even work.
12
Replying to @TheGingerBill
Tried it out on my Macbook Pro M5 with 48GB of RAM - it lagged. Worse experience than Macromedia Fireworks 8 on my Pentium 3 with 128MB RAM back in the days.
30
Replying to @manuelmaly
Interesting enough, it's smart enough to not use the build-in tool after this point and went for the CLI tool.
1
13
Replying to @mitsuhiko
I'd argue: No. Quality isn't slop - no matter what or who produced it.
86
Et tu, Opus 5.5?
1
26
Replying to @jedfrankowski @michael_kove @mattepstein
Open AI models are really bad atm. I really love my Pi agent, but it's so bad that I went back to claude. Results are much better. xitter.cf/angular2react/status/2…
angular2react@angular2reactSep 19
Replying to @angular2react
SOTA models got me so frustrated, I even wrote a whole article about how few trust I have in them xitter.cf/angular2react/status/2…
1
26
Replying to @michael_kove @mattepstein
AI does write a lot of slop. Opus 5.5 does a lot better than OpenAI models, but I've yet to come across an implementation I can take as is.
1
1
14
Replying to @ThePrimeagen
NOt here. Don't jynx it for me. I'm not really over Sol yet. Can't lose Opus 5.5 so soon.
209
Replying to @michael_kove
Yes, agreed. The more complex the system, the more it's tailored to huge corps.
1
5
Replying to @michael_kove
Got it. It's still off - but not by much. But at that level it slowly gets interesting to start to work with tax-write-offs. I think that's the true purpose of our mad tax laws - subsidizing the tax accountants and their tricks.
1
1
13
Replying to @michael_kove
...that can't be true. You don't pay income tax for your first 13k ChatGPT calculation looks more realistic. Although, if you just start out or do it as side-business, you'd get a SVS exception and only pay the accident insurance.
1
18
Load more