Saul Canina

Meta Ads | Google Ads | TikTok Ads | Shopify | Klaviyo. We help ecommerce brands scale their business building an ecosystem to drive more profit, clients etc

[Saul Canina @CaninaSaul]

@finkd Difference between Muse Wonder and Muse Wonder Full?

Really impressive release

[Saul Canina @CaninaSaul]

@JamesMalsawm Me too, addicted to Opus 5.5 billions tokens later 😅

[Saul Canina @CaninaSaul]

@thsottiaux I’m the only one who set the 1M context window on Codex?

[Saul Canina @CaninaSaul]

@thsottiaux Please more usage Tibo 🙏

[Saul Canina @CaninaSaul]

@ConnorTalksAI Opus 5.5

[Saul Canina @CaninaSaul]

@Im_IrushiK Since day one , burned billions of tokens. This model is still amazing and a pleasure to work with.

[Saul Canina @CaninaSaul]

@DenisCerednice1 Orca and T3 Code

[Saul Canina @CaninaSaul]

@DenisCerednice1 By the way T3 Code is amazing

[Saul Canina @CaninaSaul]

@DenisCerednice1 Claude Max 5x and GPT Pro 100, thinking about Claude Max 20x seriously.

[Saul Canina @CaninaSaul]

@AnkanXplorer No way 😅
OpenAI models are great for reviews and find bugs. The amount of work you can do with Opus 5.5 is absolutely insane. I use both on 100$ plan but I do the majority of my work in Claude. You need to try.

[Saul Canina @CaninaSaul]

@JamesMalsawm Sonnet and Haiku are huge upgrades, but I still didn’t find a reason to not use Opus as much as I can. It’s damn good , as orchestrator, ultracode, with Fable. Insane.

[Saul Canina @CaninaSaul]

@giadotai Opus by far for building. Astra as reviewer, pretty good at finding bugs/issues. The main problem with Astra is usage disappears and slow af.

[Saul Canina @CaninaSaul]

@Da7_Tech Claude and Codex both 100$ plan. Claude runs my business , Codex mostly reviews Claude’s work. I try but I can’t do the same stuff I do on Claude with Codex but as reviewer is great.

[Saul Canina @CaninaSaul]

@SeHozaifa I think you wrote the opposite way.
6.1 Sol is a good and cheaper model but Opus 5.5 plays in a different league.

[Saul Canina @CaninaSaul]

@karanC_12 The real question, Opus 6 or Fable 5.5 ?

[Saul Canina @CaninaSaul]

@ChatGPT Please ship Dots in Italy 🙏

[Saul Canina @CaninaSaul]

@Kappaemmedev I tested but too limited in my experience and cancel my Grok sub.
I still prefer Routines on Claude or similar feature on Codex, way more freedom, flexibility, MCPs, environments etc.

[Saul Canina @CaninaSaul]

@kimmonismus It’s not accurate based on my usage, I have both on the 100$ plan.
Last 30 days: Claude 38x around 3800$ and GPT 29x around 2900$ , maybe I can squeeze GPT a little bit more.

[Saul Canina @CaninaSaul]

@thsottiaux @makeamarkery A faster 6.1 Sol please Tibo 🙏
And better ui in the CLI regarding what subagents are doing

[Saul Canina @CaninaSaul]

@bcherny @Yuchenj_UW Tried recently and great upgrade regarding speed and overall ui/ux but my loyalty goes to the TUI, it’s really amazing.

Things could be improved:
- Better computer use
- Better ui understanding of what subagents are doing similar to the tui that I love it
- Speed similar to CLI