Christopher

Workout fanatic • Principal Engineer • YouTuber • @BigBearTech

[Christopher @dragonfire1119]

@marcelpociot @getpolyscope Tried it the other day, and it seems to use a lot of tokens. I also didn't see a way to clear the chat.

[Christopher @dragonfire1119]

@trq212 @vidhyarang @claudeai @amorriscode When I try to look at usage I get `Rate limited. Please try again later.`. 🤔

[Christopher @dragonfire1119]

@heyandras Are you using the web app/cloud for this or your computer?

[Christopher @dragonfire1119]

@ivanfioravanti Ghostty use to use iTerm but Ghostty is so much more snappy and not bloated.

[Christopher @dragonfire1119]

@RadicDavydov @heyandras Antigravity is proxying all requests to the API providers, so it has its own visibility into my code and prompts. That means it should have a clear, product‑specific policy on logging and retention at the Antigravity layer, not just rely on whatever the underlying AI vendors do.

[Christopher @dragonfire1119]

@RadicDavydov @heyandras * Retention/deletion is only described at a high level in the global policies; there’s no clear description of how long Antigravity keeps my data or how to fully purge it.

[Christopher @dragonfire1119]

@RadicDavydov @heyandras * The data‑sharing language is inherited from broad Google policies, and it’s not clear how it applies to source code and internal repos inside the IDE.

[Christopher @dragonfire1119]

@RadicDavydov @heyandras Genuine question: how are people trusting Google Antigravity when the TOS and privacy policy are this vague?

[Christopher @dragonfire1119]

@___Atin___ @aidenybai Which plan is this?

[Christopher @dragonfire1119]

@spikedoanz @aidenybai Which plan is this?

[Christopher @dragonfire1119]

This was just one test, but it made me hesitate about using GPT 5.2 for certain visual tasks. The benchmarks show it should excel here, so maybe I hit an edge case? Curious if others are seeing similar inconsistencies or if this was a fluke.

[Christopher @dragonfire1119]

Ran an interesting test today with Claude 4.5 Sonnet, Gemini 3 Pro, and GPT 5.2. Asked them to spot a name spelling error in an image. Claude caught it, Gemini caught it, but GPT 5.2 missed it. Wasn’t expecting that given the benchmark hype.

[Christopher @dragonfire1119]

Wow, GPT‑5.2 is a big step up from GPT‑5.1 — @sama and the team have made huge progress.

[Christopher @dragonfire1119]

@nunowar @pestphp I've looked into this more. The issue is caused by the .prose class injected by the Perplexity theme. If you use the Comet theme in light mode, the problem appears, but in dark mode it looks good and is readable.

[Christopher @dragonfire1119]

Pest PHP has a styling issue on the docs. @pestphp 🧐

[Christopher @dragonfire1119]

@nunowar @pestphp Hello I'm using Perplexity Comet. Let me know if you need more details.

[Christopher @dragonfire1119]

@marctuinier People would figure out how to block the ads, then you would be giving free tokens away, just like YouTube, with people justifying blocking ads.

[Christopher @dragonfire1119]

Anybody else noticing Gemini 3 just gets errors on Github Copilot now? Clicking try again x2 lets it work a bit more than another error. @Github @GoogleDeepMind Claude is working...

[Christopher @dragonfire1119]

Wow, Copilot CLI has improved a lot—it's been rock solid.