JayDax Computer Information Centre

Artificial Intelligence

There are a variety of AI platforms available to the public. Many you will have heard of:

• ChatGPT
• Gemini
• Claude
• Grok
• Perplexity

Each of these can be used free or offer a more comprehensive paid version. If you use them frequently a paid subscription may be worthwhile. Expect to pay about £20 per month for a 'Pro' model which offers significant advantages to the free version.

What can an AI do to help?

For General Internet Explorers

• Fact-Checking & Data Synthesis: Quickly verify viral social media claims, extract raw statistics from long reports, and summarize dense online articles.
• Information Retrieval: Bypass cluttered, ad-heavy search engine results to get direct, concise answers to complex research questions.
• Language & Tone Translation: Adapt text instantly to match a specific platform's audience—whether it's professional web copy, a formal email, or a short, engaging tweet under 280 characters.
• Free advice: You can ask many of the usual AIs for advice on anything but beware, it's worth checking more than one AI because, like humans, the advice may not be accurate.

For Web Designers & Developers

• Code Assistance & Troubleshooting: Quickly generate clean HTML, CSS, or JavaScript snippets, and rapidly debug broken layout structures or navigation menus.
• Content Layout & Wireframing: Brainstorm site structures, responsive design media queries, and accessible color schemes based on a specific brand image.
• Mockup Data: Instantly populate test databases or website wireframes with realistic placeholder text, user profiles, or product descriptions.
• Keeping up-to-date: Instantly summarize massive new documentation, API changes, or framework updates without reading through hundreds of forum pages.

For Authors

• Brainstorming Partner: Break through writer's block by generating plot ideas, character concepts, or alternative chapter outlines.
• Editing & Refinement: Instantly spot passive voice, suggest structural improvements, and check spelling or grammar without losing your unique voice.
• Formatting & Metadata: Quickly format text styles, generate compelling book descriptions, and create relevant SEO keywords for digital publishing.

The general perception is that creating and using an AI requires a massive server farm consuming enormous amounts of power to match the creativity of the human brain. While that scale of energy is required to build an AGI (Artificial General Intelligence) or an ASI (Artificial Super Intelligence), it isn't true for day-to-day use.

The primary reason those top platforms require so much infrastructure is that they aren't catering to one person—they are processing requests for millions of people simultaneously. If only one person needs help, a pre-trained Large Language Model (LLM) can easily be run locally on a modern home computer, using no more power than a standard video game.

"The development... has been done already"

To make this clearer, when a user runs an AI at home, they are downloading a pre-trained model. The billions of dollars and massive energy spikes happened during that initial training phase in a corporate data center. Running it at home just uses the finished "weights" (the brain's map), which is incredibly lightweight by comparison.

"Run one on a modern home computer"

While true, it's worth noting that a home computer usually runs a compressed or scaled-down version (often called a "quantized" model) rather than the exact, full-sized behemoth running on a commercial server farm. A home PC can run a model with 8 billion to 70 billion parameters beautifully, whereas the commercial top 5 use models with hundreds of billions or even trillions of parameters.