Did OpenAI finally fix ChatGPT's attitude?

Hi ,

OpenAI just released GPT-5.3 Instant, an update to ChatGPT's most-used model.

More accurate answers. Better web search results. Fewer unnecessary refusals and disclaimers. Smoother conversational flow.

Available now in ChatGPT and API.

First let's build your first-principles solution design prompt and celebrate this week's community wins. Then what GPT-5.3 Instant actually improves.

🔥 Prompt of the Day 🔥

First-Principles Solution Design: Use ChatGPT or Claude

Design solutions that solve core problems with minimum complexity and maximum impact.

The Prompt:

"You are an Expert Product Strategist and Efficiency Engineer.

Using the first-principles leverage point identified for [BIG_PROBLEM], design a solution that is [EFFICIENT], [EFFECTIVE], and [ELEGANT].

This solution must solve the core issue for [TARGET_GROUP] with the absolute minimum amount of complexity and the maximum possible impact.

Outline:

• The architecture of this solution

• The user experience

• Why this specific approach is the most logical and sustainable way to address the problem for the long term

If anything is unclear or you need additional details to improve your response, please ask me for clarification."

Variables:

• BIG_PROBLEM: The core issue you're solving

• TARGET_GROUP: Who this solves for

• EFFICIENT: Resource constraints

• EFFECTIVE: Success metrics

• ELEGANT: Simplicity requirements

🎉 Win Wednesday 🎉

This week's wins from the community.

Jeremy LeRay got OpenClaw running.

His words: "Got OpenClaw running, now to optimize how to work effectively with 'Nick'."

Named his AI agent. Now optimizing workflow with it.

Why this matters:

Getting tool installed is step one. Optimizing how to work with it is where value comes from.

Jeremy's not just running OpenClaw. He's building relationship with AI agent. Figuring out how to delegate effectively.

Charles Janoah Policarpio is considering hardware upgrade for AI work.

His words: "feeling very strongly to buy a Macbook to use Claude to its limit"

Why this matters:

Tools matter. Right hardware unlocks capabilities that wrong hardware blocks.

Charles recognizes that maximizing Claude requires right environment. Willing to invest in setup that removes limitations.

Robert Lee built custom MCP for consulting practice.

His words: "Created an MCP that launches inside Claude and fully integrates our consulting practice and allows for our members to essentially get walkthrough of the processes to build a business system from scratch. The most beautiful part is I can now focus on what I do best: help them develop their business intuition and logic so they can change the rules of how they work. It's ******* elegant."

What he built:

MCP that launches inside Claude. Fully integrates consulting practice. Members get walkthrough of processes to build business system from scratch.

Result: He focuses on developing their business intuition and logic instead of explaining processes.

Why this matters:

Robert didn't just use AI. He built infrastructure that scales his expertise.

Clients get process walkthrough through AI. Robert focuses on high-value work: developing their intuition and judgment.

AI handles documentation and process. Human handles strategy and wisdom.

That's how you use AI to multiply yourself without diluting what makes you valuable.

Sondra Verva used Claude Cowork for first time.

Her words: "I used Claude Cowork for the first time last night and it organized 520 download files in 10 minutes for me while I did other work... My first real Agentic work inside my own computer! I'm addicted to Claude now... And I managed to get Claude Code installed with VS Studio... Testing some ideas"

What she did:

Claude Cowork organized 520 download files in 10 minutes while she did other work.

First real agentic work inside her own computer.

Got Claude Code installed with VS Studio. Testing ideas.

Why this matters:

She experienced agentic AI for first time. AI working independently while she focused elsewhere.

520 files organized in 10 minutes. That's not assistance. That's automation.

She's not just learning tools. She's testing ideas. That's when real innovation happens.

What did you ship this week?

New system? First sale? Tool mastered?

Reply and share what you built.

Did You Know?

Waymo now provides over a hundred thousand self-driving rides per week across U.S. cities, while Baidu's Apollo Go robotaxi fleet serves numerous cities across China - autonomous vehicles have moved from experimental curiosity to everyday transportation.

🗞️ Breaking AI News 🗞️

OpenAI Releases GPT-5.3 Instant

OpenAI announced GPT-5.3 Instant, update to ChatGPT's most-used model.

Makes everyday conversations more consistently helpful and fluid.

Available now in ChatGPT and API.

What Changed

Previous GPT-5.2 Instant: Would sometimes refuse questions it should answer safely. Overly cautious or preachy responses, especially around sensitive topics.

GPT-5.3 Instant: Significantly reduces unnecessary refusals. Tones down overly defensive or moralizing preambles. Provides useful answers directly without unnecessary caveats.

Result: Fewer dead ends. More directly helpful answers.

Better Web Search Synthesis

GPT-5.3 Instant improves quality of answers when information comes from web.

More effectively balances what it finds online with its own knowledge and reasoning.

Example: Uses existing understanding to contextualize recent news rather than simply summarizing search results.

Less likely to overindex on web results. Previously could lead to long lists of links or loosely connected information.

Stronger job recognizing subtext of questions. Surfaces most important information upfront.

Result: More relevant and immediately usable answers without sacrificing speed or tone.

Smoother Conversational Style

GPT-5.2 Instant's tone could sometimes feel overbearing. Made unwarranted assumptions about user intent or emotions.

GPT-5.3 Instant: More focused yet natural conversational style. Cuts back on unnecessary proclamations. Fewer phrases like "Stop. Take a breath."

Working to keep ChatGPT's personality more consistent across conversations and updates.

Users can still adjust model's response tone (warmth, enthusiasm) within settings.

More Accurate Responses

Delivers more factual responses. Reduced hallucinations across wide range of topics.

OpenAI measured accuracy using two internal evaluations:

Higher-stakes domains: Medicine, law, finance.

User-flagged errors: De-identified ChatGPT conversations users flagged as factual errors.

Results:

Higher-stakes evaluation: 26.8% reduction in hallucinations when using web. 19.7% reduction when relying only on internal knowledge.

User-feedback evaluation: 22.5% reduction in hallucinations with web use. 9.6% reduction without web access.

Stronger Writing

Better at helping write resonant, imaginative, immersive prose.

Whether drafting fiction, refining passage, or exploring new ideas.

Moves more fluidly between practical tasks and expressive writing without losing clarity or coherence.

Limitations

Non-English languages: Response style in some languages (Japanese, Korean) can sound stilted or overly literal. Improving tone and naturalness across languages remains ongoing focus.

Tone: While response tone should feel smoother, OpenAI continuing to monitor feedback and improve while expanding customization options.

Availability

GPT-5.3 Instant: Available now to all ChatGPT users. Available to developers in API as 'gpt-5.3-chat-latest.'

Updates to Thinking and Pro modes: Follow soon.

GPT-5.2 Instant: Remains available for three months for paid users in model picker under Legacy Models section. Will be retired June 3, 2026.

Why This Matters

This is refinement, not revolution.

OpenAI fixing annoying parts of ChatGPT. Unnecessary refusals. Overly cautious responses. Poor web search synthesis. Hallucinations.

For users: ChatGPT feels more helpful. Less friction. Fewer dead ends.

For developers: API gets more reliable model. Fewer hallucinations matters for production applications.

For competitive landscape: OpenAI maintaining ChatGPT quality while Anthropic, Google compete on features.

What This Means

If you use ChatGPT, your conversations just got smoother automatically.

If you're developer using API, update to 'gpt-5.3-chat-latest' for better accuracy.

If you build products on ChatGPT, reduced hallucinations means more reliable outputs.

Small improvements compound. GPT-5.3 Instant doesn't add new capabilities. It makes existing capabilities work better.

Over to You...

How many hours per week would you save if ChatGPT stopped giving unnecessary caveats before answering?

Hit reply and share what becomes possible.

To efficiency gains,

Jeff J. Hunter 
Founder, AI Persona Method | TheTip.ai

P.S. If you’re serious about turning AI agents into a premium consulting offer - book your AI Certified Consultant strategy call👉 here.

NEW: “AI Money Group” to Learn to Make Money with AI