Skip to content
How I AIHow I AI

Claude is BACK with Opus 5.5

I’ve been off Claude for months. Not because it got dumb, but because it got annoying. The rambling, the hedging, the preachy little disclaimers on tasks that didn’t need them. I moved most of my daily work to Codex and I didn’t miss it. Then Anthropic shipped Opus 5.5: 40% cheaper than Opus 5, faster, and with what they’re calling a fundamentally different alignment approach. I ran it for a week across real work, including four long-running agentic tasks, a full ChatPRD homepage redesign, an SVG benchmark, and one very firm refusal, and I’m ready to give you the honest verdict. There’s a lot to like. There are still two things that drive me a little crazy. And there’s one capability I genuinely wasn’t expecting. *What you’ll learn:* 1. Why I walked away from Claude entirely, and what it took for me to come back 2. The real cost math on Opus 5.5 and why pricing matters more for agentic work than single prompts 3. What happened when I ran four long-running agentic tasks, including one that tried to manipulate Claude mid-run 4. Why Opus 5.5 is now my go-to for frontend prototyping, and where it still lets me down 5. The one capability I genuinely didn’t see coming, and no other model in my stack can match it 6. The moment Opus 5.5 told me flat-out no, and what that says about where Anthropic’s safety posture actually lands in practice 7. Where Codex still wins, and how I’m splitting my model stack after a full week of testing *In this episode:* (00:00) Why I stopped using Claude (01:02) What Anthropic says Opus 5.5 is (01:54) Cost, speed, and benchmark overview (03:20) Safety, alignment, and the cybersecurity limits (05:02) How I AI bench (05:39) Voice test: is it actually not annoying? (07:54) Long-running agentic task results (10:50) Frontend prototyping (17:23) Writing voice and email (19:41) SVG illustrations (20:46) Video editing (21:42) My verdict: what it’s good at, what it still isn’t *Tools referenced:* • Claude Opus 5.5: https://www.anthropic.com/claude-opus-5-5 • ElevenLabs MCP connector: https://elevenlabs.io/mcp • Codex (OpenAI): https://openai.com/codex *Where to find Claire Vo:* ChatPRD: https://www.chatprd.ai/ Website: https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ X: https://x.com/clairevo _Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email jordan@penname.co._

Claire Vohost
Sep 22, 202624mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Opus 5.5 makes Claude usable again, especially for design

  1. Claire returns to Claude after months away because Opus 5.5 finally feels minimally annoying and more concise in conversation.
  2. Anthropic positions Opus 5.5 as faster and ~40% cheaper than Opus 5 with strong benchmarks; Claire agrees it’s zippier but notes quiet long turns can worsen perceived speed.
  3. In real work, Opus 5.5 completes long-running agent tasks (25–82 steps) and demonstrates strong instruction-following, edge-case spotting, and prompt-injection resistance.
  4. Opus 5.5 excels at frontend prototyping and produces strong SaaS/devtools UI concepts, though it still shows “slop” in copy, alignment, and consumer-style aesthetics.
  5. It shines at SVG illustration generation but underperforms in automated short-form video editing, and Claire still prefers Codex for computer-use workflows and overall harness UX.

IDEAS WORTH REMEMBERING

5 ideas

Opus 5.5 fixes the “annoying Claude voice” enough to bring Claude back into her workflow.

Claire’s main blocker wasn’t capability but interaction quality: previous Claude versions felt rambling and frustrating to use. Opus 5.5 feels more direct and readable, closer to a “GPT-like” succinctness, which makes it viable again for daily work.

The big UX tradeoff is perceived latency: less narration, more silence.

Anthropic markets Opus 5.5 as ~Fable-level performance with lower cost and higher speed; Claire confirms it feels “zippier” than Opus 5. However, longer turns can feel slow because the model stays silent for minutes, increasing perceived latency even if throughput is good.

It’s reliable for long-running agentic workflows and shows strong instruction-following and injection resistance.

Across inbox triage, backend feature building, long-running research, and computer-use style tasks, all runs completed successfully with 25–82 steps per prompt. She notes it followed instructions well, found outliers/edge cases, and resisted a prompt injection during inbox triage.

Best-in-class for frontend prototyping (especially SaaS/devtools), but still needs design direction and cleanup.

Claire’s strongest praise is for UI prototyping: it one-shots complex, interactive prototypes and produces visually strong SaaS layouts (spacing, color, gradients, dashboards). Limitations remain: occasional “slop” copy, alignment issues, over-dense “chaos rain” layouts, and weak consumer-app aesthetics (Claude’s beige/orange tendencies).

SVG illustration generation is a surprising flagship strength for Opus 5.5.

She found Opus 5.5 uniquely strong at generating coherent, cute SVG illustrations with consistent style and fewer structural bugs than other models. This becomes a standout “new” use case she recommends trying immediately.

WORDS WORTH SAVING

5 quotes

I stopped using Claude 'cause it was annoying. Annoying.

Claire Vo

But I don't care about that. I care is it annoying? And guess what guys? We did it. It is not annoying anymore, or at least it's minimally annoying.

Claire Vo

Claude is gonna scold you. Claude is kind of square. Claude is not gonna drink with you in a field behind your friend's house. Like Claude's not a party boy.

Claire Vo

I don't wanna be told no by my AI.

Claire Vo

So just, like, going back to the top of what I think about Opus 5.5, one, it is not annoying.

Claire Vo

Why Claire stopped using Claude (annoyance/verbosity)Opus 5.5 positioning: cost, speed, benchmark claimsSafety, alignment, and cybersecurity guardrailsLong-running agentic task performance (inbox triage, research, coding, computer use)Frontend prototyping strengths and design limitationsEmail voice mimicry and assistant behavior (scolding/refusals)SVG illustration generation vs. video-editing weakness

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.