AI Newsletter 1_23_26

3 views
Skip to first unread message

Caleb Bryant

unread,
Jan 23, 2026, 9:51:46 AMJan 23
to IT HootClub
OwlAtComputer.png
IT HootClub — AI Community Newsletter
Hands-on. Career-focused. Future-ready.
Issued 2026-01-23

This email is the full newsletter. For a printable copy, click here to download the PDF.

Join the Dialogue
Share thoughts and questions on our non-invasive discussion page. No email blasts; posts appear after moderator approval.

Open the Dialogue

The Learning Loop
DEFINITION Fine-tuning
Fine-tuning is like teaching a dog new tricks. You start with a dog (a pre-trained AI model) that already knows basic commands (general knowledge). Instead of starting from scratch, you show it specific commands (new data) and reward it for getting them right. Over time, it learns to perform those specific tasks better than if you'd taught it from zero.
Source:
TIP Prompt Segmentation
Break down complex requests into smaller, sequential prompts. This allows the AI to focus on one aspect at a time, reducing errors and improving the quality of the final output. Imagine guiding a painter, step by step, instead of throwing all the instructions at once.
Source:
TOOL Descript
Descript is a collaborative audio and video editor that uses AI to transcribe, edit, and enhance your recordings. Edit audio and video like you would a text document!
Under the Hood
Greetings! Last week I spent some time discussing the migration of using Gemini on https://gemini.google.com/ versus using Google AI Studio (https://aistudio.google.com). Really, they are both the same thing, except Google AI Studio is where you obtain the API Key. The API Key is the "key" (a string of unique identifiable numbers) that Google assigns to you. You insert this key into your code so that your program can call the AI model for whatever function it is you are wanting to use the AI model for. The difference between the two is that Google AI Studio is more for development purposes. The gemini.google.com site charges you for using Gemini directly through that site. Google AI Studio charges you based on the specific usage of the API Key that you're given and how data is processed when it "calls" Gemini to action. So, what are the costs for this? AI API pricing typically works on a pay-per-use model based on tokens. Here's how it generally breaks down: Tokens are the billing unit. A token is roughly 3-4 characters or about ¾ of a word in English. Both your input (the prompt you send) and the output (the model's response) are measured in tokens, Input vs. output pricing. Most providers charge separately for input and output tokens, with output tokens usually costing more (often 3-5x more). This is because generating text is more computationally expensive than processing it. Pricing varies by model. More capable models cost more. For example, a flagship model costs significantly more per token than a smaller, faster, or older model. You choose based on your task complexity and budget. Costs are usually quoted per million tokens. As a rough sense: Smaller/faster models: ~$0.25–$1 per million input tokens, Mid-tier models: ~$3–$5 per million input tokens, Flagship models: ~$10–$20+ per million input tokens, Other factors that can affect cost: Context window usage — some providers charge more for using extended context lengths, Fine-tuned models — custom models often have higher rates, Cached/prompt caching — some APIs offer discounts when you reuse the same prompt prefix repeatedly, Batch processing — sending requests in bulk can sometimes reduce costs. Practical example: If you send a 500-token prompt and receive a 1,000-token response using a mid-tier model, you might pay something like $0.002–$0.005 for that single request. Costs add up with volume. For something like a weekly newsletter that I have done, the API Key is called once a week. The cost is negligible, literally like a quarter, at most, each newsletter. The volume is where it can start to get pricey and this is where a business operating a website has to pay attention. If you are thinking about implementing an API Key to become more automatic in your AI usage, I would encourage you to do so. Technically, you could get surprised by a large bill, however each of these model providers have pretty fail safe procedures in place to keep you notified and your account capped at a specific cost. You can easily monitor your usage and billing as well. AI model usage cost is very interesting right now. There are many ways for a business to structure this so that it fits with their revenue model. It should be noted that the scenario I just covered is a typical scenario. Here, I am mentioning Google, specifically at times, because that is just one frontier model that I have used. Obtaining API keys through other models is a different, however similar, process. That's all for my AI automation journey this week. Next week we get back to the migration to Google AI Studio and what I experienced there. Bye for now!
Lift-Off

“The pursuit of science is a grand adventure, fueled by curiosity and a willingness to be wrong.”

— Vera Rubin — Vera Rubin was an American astronomer who pioneered work on galaxy rotation rates. Her groundbreaking research provided evidence of dark matter. She received numerous awards and honors, including the National Medal of Science.
The Nest Jest
The_Nest_Jest.png
Upcoming Events
Event 01/29/2026 6:00 PM
AI Coding Tools Demo Night - Joint with MadAI! — Madison College, Rooms 309 & 311, Madison, WI
**Join us for a special hands-on “AI Coding Tools Demo Night” on Thursday, January 29th, 2026 at 6:00pm**, hosted at the Madison College Truax campus. Over the course of two hours, **3–4 presenters will each use a different AI-assisted coding tool**—Copilot, Cursor, Claude Code, Gemini, Codex, etc—to build the same small app from scratch in about **20 minutes each**, from first prompt all the way through deployment. This is a great chance to see how these tools actually behave in real time and t [Group: madison-ai]
Event 02/10/2026 10:00 AM
2 Days Virtual Hands on Workshop : Multi-Agent AI Systems with Copilot Studio — Online
**2-Day Hands-on Workshop: Crafting Multi-Agent AI Systems & Autonomous Agents with Copilot Studio, Power Platform, Dynamics 365, MCP, ChatGPT, DeepSeek, & Azure AI Foundry** **February 10-11, 2025** **9:00 AM – 5:00 PM (Central)** **Registration Link:** https://dynamicscommunities.com/learning-lab-microsoft-copilot-studio/ **Copilot Studio** is a graphical, low-code tool for both creating an agent—including building automation with Power Automate—and extending a Microsoft 365 Copilot with you [Group: artificialintelligenceandmachinelearning]
Event 02/11/2026 5:15 PM
February Global AI Milwaukee User Group Meeting — Wctc, 800 Main Street, Pewaukee, WI
Agenda Includes: * Networking / Food * Brief Introduction / Discuss Group Business * Featured Speakers Featured Speaker: **TBD** Topic: **TBD** **Featured Sponsors** TBD [Location Sponsor] TBD [Food Sponsor] [Group: global-ai_milwaukee]
Event 2/12/2026
NMDSI Presents Elevating and Accelerating AI Excellence in Southeast Wisconsin — Milwaukee, WI
Corporate Innovators - Northwestern Mutual Commons - Grand Hall, 805 East Mason St, Milwaukee, WI 53202
Source: Choose MKE Tech
In the News
News 2026-01-21 — TechCrunch
Irony alert: Hallucinated citations found in papers from NeurIPS, the prestigious AI conference
Research from GPTZero indicates that prestigious conferences are facing challenges related to AI-generated content, including "hallucinated" or fabricated citations within submitted papers. This highlights the difficulty of maintaining integrity in academic publications amidst the rise of AI-generated slop. read more
News 2026-01-21 — Wired
The US and China Are Collaborating More Closely on AI Than You Think
WIRED analyzed over 5,000 papers from NeurIPS using OpenAI’s Codex to understand the areas where the US and China collaborate on AI research. The analysis sheds light on the extent of US-China cooperation in the field of artificial intelligence research. read more
News 2026-01-21 — Wired
Thousands of Companies Are Driving China’s AI Boom. A Government Registry Tracks Them All
The Cyberspace Administration of China has inadvertently created a guide to the country's burgeoning AI industry. This registry provides insights into the numerous companies contributing to China's AI revolution. read more
This newsletter was assembled with an AI-assisted workflow prototype for internal preview.
Reply all
Reply to author
Forward
0 new messages