Boost Productivity: 5 Ways to Use Claude, Perplexity, Gemini

Try Claude 3 Opus: tight reasoning, sharper assets

• What’s new: Anthropic shipped small model and interface tweaks. They call it incremental, but those increments matter on longer tasks.
• Early reports: better accuracy on multi-step reasoning, cleaner slide and HTML output, and mixed results on quick copy jobs.
• How to try: open Claude, pick Opus (latest) from the model dropdown, and rerun a work prompt you know well.

Why this matters

• You get fewer follow-ups and clarifying edits when the model holds a long chain of reasoning.
• That saves time: fewer back-and-forths, faster approvals, less “Did you mean…?” from the client.

Mini snapshot

• A product manager I spoke with ran a 10-slide brief through Opus. She told me the first draft needed “far fewer edits,” and her review time dropped noticeably.

Claude Design: a mini Figma/Canva that actually asks questions

Meet claude.ai/design — an experimental, browser-based creative studio powered by Opus. Think of it as a quick visual lab inside Claude.

Three things it does well:

1. Presentation builder
   • Give it an outline or script. It asks clarifying questions, then drafts slides with imagery, a color palette, and speaker notes. The interrogation up front makes drafts feel sharper.
2. Motion graphics
   • Paste a table or a screenshot. Ask for a short animation. Seconds later you have an export-ready MP4.
3. Web & email templates
   • Prompt “one-page product site” or “launch newsletter,” tweak visually, then download HTML.

Why this matters

• Designers who dread late “just whip up a deck” requests will like this. The tool nudges you toward better inputs, so the outputs need fewer fixes.

Google’s Gemini on macOS: quick access, a few caveats

Google released a Gemini desktop app for macOS. Useful bits:

• Fast launcher: hold Option + Space to open a prompt box anywhere.
• Image generation: Imagen 2 is integrated for on-device image work.
• Seamless copy-paste: move results into any Mac app with no fuss.

Notable limits

• The macOS app still lacks Gems and Notebooks (two web-only power features).
• Editing controls for generated text and images are lighter than the browser version.

Personalisation angle

If you let Gemini access Google Photos, Imagen 2 tailors images to your style. The results can be very on-brand — or delightfully weird.

Gemini in Chrome: save time with “/skills”

Gemini’s Chrome sidebar now supports /skills — saved, reusable prompts.

• Example: type /summarise under a research article, and your canned summarizer runs automatically.
• Rollout is staged, so you might not see it yet. If you do, set up one skill for the task you repeat daily.

NotebookLM: prettier notebooks, now on free tier

• You can add custom cover art and rewrite notebook descriptions for cleaner scanning.
• Chats from Gemini that reference a Notebook now appear as sources without counting toward your Notebook limits.
• NotebookLM data (sources and chats) stays out of Google’s model training.
• The Gemini-Notebook integration is rolling out to free accounts globally.

Perplexity Personal Computer: a local agent that watches folders

Perplexity’s “Personal Computer” turns a Mac mini into a small, always-on agent. It can:

• Read and write local files.
• Trigger reminders, emails, and notes.
• Watch folders and fire workflows when files appear.

Practical examples

• Drop a raw podcast recording in “To Process.” The agent transcribes it, drafts show notes, and moves final files.
• Save a YouTube outline. The agent creates a Keynote deck and social captions.
• Screenshot a receipt. It files the image under this month’s expenses.

Status: closed beta with a few visual glitches. But it’s a clear hint at what local agents can do.

What to try first

1. Open Claude 3 Opus (latest) and rerun one complex prompt you use weekly. Compare drafts.
2. Build a 3‑slide deck in claude.ai/design and time the process.
3. On macOS, press Option + Space and use Gemini as a system-wide assistant for a day.
4. In Chrome, create one /skill for a repeatable workflow.
5. Add a cover image to a NotebookLM doc and notice how scanning improves.

If you use AI for multi-step work, these updates shave time and friction. One sentence takeaway: rerun your real prompts and pick the tool that actually saves you edits.

Want a friendly place to learn how to use these tools—step by step? Check out tixu.ai — beginner-friendly AI courses and hands-on labs to get you productive faster: tixu.ai