Get Far Better ChatGPT Answers Using These Three Simple Tricks ✨

Welcome to another edition of Horizon AI,

There are improvements to how you prompt ChatGPT and to forcing the "thinking" model to specify exactly how long you want the answer and to structure complex requests, which dramatically improve accuracy, relevance, and usefulness.

Let’s jump into it!

Read Time: 4.5 min

Here's what's new today in the Horizon AI

• Google Executives Crush Gemini Ad Rumors 🚫

• Anthropic Brings Claude Directly into Slack

• AI Tutorial: Get Far Better ChatGPT Answers Using Three Simple Tricks

• AI Tools to check out

• AI Findings/Resources

• The Latest in AI and Tech 💡

AI News

GOOGLE

Google Executives Crush Gemini Ad Rumors

Following viral reports claiming that Google was pitching advertisers on Gemini ad slots for 2026, the tech giant has issued a swift, categorical denial. Senior Google executives have taken to social media to dispel speculation, clarifying that the standalone Gemini app is not currently being prepared for monetization, despite anonymous industry sources.

Details:

• Dan Taylor, Google's VP of Global Ads, publicly addressed the rumors on X (formerly Twitter), stating the original story was based on "uninformed, anonymous sources who are making inaccurate claims."

• Taylor provided a definitive statement regarding the company's current roadmap: "There are no ads in the Gemini app, and there are no current plans to change that."

• The denial highlights a critical distinction in Google's strategy. While the company is actively placing ads in "AI Overviews" (the AI summaries within standard Google Search) and testing them in "AI Mode," it is drawing a clear line between those search features and the standalone Gemini chatbot app.

• The rapid response underscores how sensitive Google is to user perception right now. With competitors like ChatGPT and Claude focusing on clean, ad-free subscription models, Google is keen to avoid the perception that its assistant is becoming just another billboard.

Source

TOGETHER WITH RAD INTEL

Missed OpenAI? The Clock Is Ticking on RAD Intel’s Round

Ground floor opportunity on predictive AI for ROI-based content.

RAD Intel is already trusted by a who’s-who of Fortune 1000 brands and leading global agencies with recurring seven-figure partnerships in place.

$50M+ raised. 10,000+ investors. Valuation up 4,900% in four years*.

Backed by Adobe and insiders from Google. Invest now.

Lock Shares

_{This is a paid advertisement for RAD Intel made pursuant to Regulation A+ offering and involves risk, including the possible loss of principal. The valuation is set by the Company and there is currently no public market for the Company's Common Stock. Nasdaq ticker “RADI” has been reserved by RAD Intel and any potential listing is subject to future regulatory approval and market conditions. Investor references reflect factual individual or institutional participation and do not imply endorsement or sponsorship by the referenced companies. Please read the offering circular and related risks at invest.radintel.ai.}

ANTHROPIC

Anthropic Brings Claude Directly into Slack

Anthropic is taking a sledgehammer to the barrier between developer communication and execution. With the launch of a new beta integration, the company is embedding "Claude Code," its specialized coding agent, directly into Slack.

Details:

• Developers can tag @Claude directly in a Slack channel to initiate coding tasks. The AI analyzes the thread's context (e.g., a bug report or a feature discussion) and takes action.

• Unlike a standard chatbot, this integration connects to the codebase. It can search repositories, analyze file structures, and propose edits based on the Slack discussion.

• The tool allows teams to go from a reported issue in chat to a generated pull request in minutes, significantly reducing the "context switching" tax developers pay when toggling between Slack and their IDEs.

• The feature is currently in beta/preview and is aimed at engineering teams looking to streamline their "ChatOps" workflows.

This move transforms the chat platform from a place where engineering teams discuss code into a place where they can actively ship code, all without leaving their conversation threads.

Source

AI Tutorial

Get Far Better ChatGPT Answers Using Three Simple Tricks

1. Force deeper reasoning: add "think hard about this"

   ChatGPT can switch to a more careful thinking mode. Adding a line like “think hard about this” or “think deeply about this” tells it to slow down and reason more. When this happens, you may see a note like “Thought for Xs” in the reply.

• You’ll get fuller, more thoughtful answers instead of quick one-liners.

  Example:

Explain the economic effects of price controls on housing supply — and think hard about this.

2. Tell ChatGPT exactly how long you want the response

   By default, ChatGPT can be verbose. If you need a short summary, a 150-word explainer, or a three-bullet list, telling it the length saves editing time and yields exactly the format you want. You can do this by adding a length instruction at the end of your prompt, e.g.: "Keep this under 100 words," "Summarize in 3 bullet points," or "Write exactly 250 words."

• Do this to get compact, usable answers that fit your audience (emails, slides, tweets).

  Example:

Summarize the main causes of the 2008 financial crisis in three bullet points. Keep each bullet under 25 words.

3. Structure complex prompts with clear sections

   When tasks are complicated, a single long paragraph confuses the model. Breaking your request into labeled sections (Context, Task, Requirements, Constraints, Output Format) gives the model clear signals and priorities.

• This gives clear, structured deliverables (slides, plans, code) aligned to your needs.

  Example:

Context: I'm presenting to product managers on data-driven prioritization.
Task: Create a 5-point slide outline explaining RICE vs. ICE prioritization.
Requirements: Use simple language, include an example for each method.
Constraints: No jargon; avoid >100 words per slide note.
Output Format: Numbered list (Slide 1, Slide 2...), each with 2–3 bullet speaker notes.

Here's an example of using these 3 simple tricks together for the best results on ChatGPT

Context: I'm preparing a 7-minute talk for startup founders about AI ethics.
Task: Write a 6-slide outline with speaker notes.
Requirements: Practical examples, one call-to-action slide, friendly tone for founders.
Constraints: Each slide note ≤ 80 words; do not use heavy academic jargon.
Output Format: Numbered slides (Slide 1: Title; Slide 2: ...), plus a 2-sentence summary.
— and think hard about this. Keep the whole response under 450 words.

AI Tools to check out

💻 RightNow: AI-powered code editor for GPU kernel development and optimization

📝 Marble: Super simple headless cms for managing your blog and media

🎶 ACE Studio 2.0: AI music workstation that brings powerful AI models into one smooth workflow.

🎥 ElevenLabs Image & Video: Generate visuals and add high-quality voiceovers, music, AI sound effects, and captions.

🤝 Delight.ai: Branded AI concierge that makes every experience personal and delightful for customers, trusted and profitable for businesses

TOGETHER WITH BRIGHT DATA

74% of Companies Are Seeing ROI from AI.

Incomplete data wastes time and stalls ROI. Bright Data connects your AI to real-time public web data so you launch faster, make confident decisions, and achieve real business growth.

Get Access

AI Findings/Resources

📈 How to make 10x ads entirely with AI and replicate any viral video.

🔮 Gemini 3 uses text, images, video, code, and audio to learn, build, and plan anything multimodally.

🧰 A collection of task-specific workflows that enhance and customize what Claude can do.

🤯 AI Video in 2023 VS 2025 (Will Smith eating spaghetti meme)

🧪 AI tools accelerate research productivity, with quality concerns.

The latest in AI and Tech

Google's AI Try-On App Doppl Gets a Shoppable Discovery Feed

Doppl, Google's experimental virtual try-on app, now offers a scrollable feed of AI-generated outfit suggestions, making it easy to discover and purchase clothes directly through the app.

Even as Gemini Shines, Experts Say AGI Is Still Far Off

Despite the buzz around powerful models like Gemini 3, researchers at NeurIPS 2025 warned that scaling up current models still doesn't deliver true understanding or cause-and-effect reasoning, casting doubt on near-term Artificial General Intelligence.

U.S. Commerce Dept. May Greenlight Nvidia H200 for China

In a major pivot for U.S. tech policy, reports suggest the Department of Commerce is poised to approve the export of Nvidia's powerful H200 chips to China, potentially legitimizing a trade that has been driven underground. Nvidia hailed the decision as a win for U.S. manufacturing and jobs, though critics argue it feeds China's AI ambitions.

Z AI Debuts Open-Source GLM-4.6V, A Vision-Capable Language Model Tool

Z AI has released GLM-4.6V, an open-source model that natively supports vision inputs, allowing users to combine text and images in the same tool, which could accelerate multimodal AI workflows and broaden what developers can build. The launch underscores growing momentum behind open-source AI, giving more people access to powerful tools without the usual cost or proprietary restrictions.

NVIDIA CEO Warns China Could "Build a Hospital in a Weekend" with Its AI Edge

The CEO of NVIDIA has issued a stark warning that China's rapidly growing AI infrastructure, now believed to account for 30% of global AI usage, gives it a significant advantage, potentially enabling massive-scale deployment (even of critical infrastructure like hospitals) with unprecedented speed.

That’s a wrap!

Thanks for sticking with us to the end! Let’s stay connected on LinkedIn and Twitter.

Not subscribed yet? Sign up here and send it to a colleague or friend!

See you in our next edition!

Gina 👩🏻‍💻