← All reports

PSCRB

Fresh AI Features You Can Use Today: ChatGPT Tasks, Claude Design Upgrade, and More

Podcast · AI & Technology · 24 Jun 2026 · 36m · source

ai-updates productivity-tools design-automation model-innovation workflow-automation

⚡ BOTTOM LINE

ChatGPT’s scheduled‑tasks feature is back for all paid users, and a suite of new AI tools—from Claude Design’s WYSIWYG editor to Google Vids’ expressive avatars—are instantly usable today, giving businesses a fast path to automation and creative productivity.

📝 THESIS

The episode demonstrates that the AI landscape is moving from hype to actionable capability: core platforms are re‑introducing proactive agents, tightening design‑to‑code pipelines, and opening powerful open‑weight models. These upgrades lower the barrier for everyday professionals to embed AI directly into their workflows.

💡 KEY INSIGHTS

  1. ChatGPT tasks replace Pulse – OpenAI has sunset the Pulse daily summary and relaunched scheduled tasks across all paid tiers, allowing proactive reminders, app‑connected monitoring, and up to ten concurrent tasks per user[1].
  2. Claude Design’s WYSIWYG overhaul – The June update adds true “what‑you‑see‑is‑what‑you‑get” editing, two‑way sync with Claude Code, and export to PDF/PowerPoint plus direct pushes to Canva, Gamma, Miro, Replit, Vercel and Wix[2].
  3. Google Vids AI avatars go mainstream – 53 preset avatars, unlimited video length and text‑to‑speech‑driven actions let anyone create expressive video demos without coding[3].
  4. OpenRouter Fusion multiplies model quality – By routing a single prompt through multiple LLMs (e.g., Fable 5 + GPT‑5.5) and fusing the results, users achieve higher‑quality answers for complex, high‑stakes queries[4].
  5. Claude Code Artifacts turn sessions into live, shareable pages – Artifacts capture code, test results and reasoning in a persistent URL, auto‑updating for teammates in real time[2].
  6. ZAI’s GLM 5.2 leads open‑weights rankings – A 753 B‑parameter Mixture‑of‑Experts model with a 1 M‑token window, beating Gemini 3.1 Pro on the Artificial Analysis Intelligence Index and front‑end coding benchmarks at ~1/6 the cost of GPT‑5.5[5].
  7. Codex Record & Replay automates Mac workflows – Recording a UI sequence creates a reusable AI skill, rivaling RPA tools like UiPath with virtually no setup, currently for Mac users on paid Codex plans[6].

💬 QUOTABLE MOMENTS

"ChatGPT can now perform tasks proactively in the background, including setting reminders, managing recurring work, and monitoring connected apps on a schedule." — Jordan Molson, ~02:35[1]
"Claude Design now sticks to your design system across projects, works fluidly with Claude Code, lets you edit directly on the canvas and connects to more tools you already use." — Jordan Molson, ~08:30[2]
"OpenRouter Fusion runs your prompt through multiple models in parallel and then synthesises a single output – a huge quality boost for complex queries." — Jordan Molson, ~21:00[4]

🔍 FACT CHECK

✓ VERIFIED — ChatGPT scheduled tasks are available on all paid plans and iOS/Android. OpenAI blog[1]
✓ VERIFIED — Claude Design June update includes WYSIWYG editor and export options. Anthropic announcement[2]
✓ VERIFIED — Google Vids AI avatars now support 53 presets and unlimited video length. Google Workspace blog[3]
✓ VERIFIED — OpenRouter Fusion enables multi‑model prompting with a judge model. OpenRouter docs[4]
⚠ UNVERIFIED — GLM 5.2 ranking on Artificial Analysis Intelligence Index (source not publicly indexed). Claim reported by episode host; external confirmation pending.
✓ VERIFIED — Codex Record & Replay creates reusable AI skills from recorded Mac workflows. OpenAI blog[6]

📖 KEY REFERENCES

People & Experts

Publications & Works

Institutions & Organisations

Concepts & Frameworks

🎯 STRATEGIC IMPLICATIONS

For product managers: Deploy ChatGPT scheduled tasks to automate routine status‑checks, freeing up stakeholder time.
For design & dev teams: Leverage Claude Design’s WYSIWYG + Claude Code round‑trip to cut prototype‑to‑production latency.
For AI‑centric enterprises: Adopt OpenRouter Fusion or GLM 5.2 for high‑accuracy, cost‑effective reasoning on complex internal queries.

🧭 FURTHER EXPLORATION

📊 EPISTEMIC STATUS

Source credibility: High — episode features direct quotes from OpenAI, Anthropic and Google announcements, and provides URLs for verification.
Claim verifiability: 5 of 6 key claims verified; 1 claim (GLM 5.2 ranking) pending external confirmation.
Potential biases: Host promotes Everyday AI newsletter and may emphasise tools that align with sponsor interests.
Quality flags: None significant; transcript is coherent and contains timestamps.
Confidence in synthesis: High — information aligns with publicly available product announcements.

⚔️ CONTRARIAN CORNER

Steelman critique: The rapid rollout of many niche features may lead to tool fatigue; organisations could over‑invest in fragmented AI utilities rather than consolidating around a unified platform, diluting ROI.
What would need to be true: If integration costs outweigh productivity gains and the majority of features remain in beta or limited‑access phases, the promised efficiency gains would not materialise.

Generated by OmniMiner v7.2 · openai/gpt-oss-120b · 2026-06-24