5 Things You Should Know in AI This Week — January 16, 2026

For you: Want these stories personalized to your role?
Try the interactive version →

This launch got a ton of attention on social media.

In theory, Cowork is meant to be Claude Code for non-technical people.

A way for AI to work with files stored on your computer.

Which takes AI out of chat mode and makes it able to complete work more autonomously.

It can read, edit, create, and organize your stored files - which means less copying and pasting for you.

But a lot of users were left underwhelmed once they tried the tool.

Turns out Cowork was built entirely with Claude Code in a week and a half.

This is impressive, but it also shows it was spun up without much strategic effort.

Here's how Claude Cowork stacks up to Claude Code function-wise:

Users were frustrated by confusing onboarding, tool connections not working, slow performance, and an interface that still felt designed for devs.

Many also felt there was too much micro-managing and not enough set-and-forget.

Cowork is still in research preview and available to Claude Max subscribers.

The signal?

The drop drew excitement because it showed where AI is headed:

• More accessible tools for people who aren't devs
• Ability to work with existing tools, documents, and file structure
• Ability to pass off more work to AI (a move towards agents)

It's early days, so there is still hope for what the tool can become.

The signal here is to start thinking beyond the chat window.

In the meantime, if you're looking for a simple tool with a friendly interface that can help with agentic work, check out Manus.

I just dropped a step-by-step guide: Get started with Manus in <15 minutes.

I don't want this newsletter to become a Google blog, but they can't stop releasing stuff...

Their latest drops:

• New AI features in Gmail
• Personal Intelligence

Here's what's new in Gmail:

• AI Inbox: Summarizes your emails daily and flags the important things
• AI Overviews: Ask questions and it pulls answers from your email history
• Writing Tools: Helps start an email, better replies, tone enhancements

As for Personal Intelligence, this release is exciting.

It'll allow Gemini to pull knowledge about you from Google apps you let it access.

It should improve personalization and memory by letting Gemini access information from Gmail, Google Photos, YouTube, Search, and other Google interaction data regardless of the app you're using.

Then it can reason and give responses based on broader info it has about you.

Personal Intelligence is intriguing, but it raises a question:

Can it prioritize info like a person or will it send you down the wrong rabbit holes?

The tire example could easily go off the rails if it overprioritizes irrelevant stuff.

The signal?

The Gmail features should be genuinely useful.

While Personal Intelligence should reduce some of the issues Gemini had...

Mainly, the feeling of talking to separate Geminis when you talk to it in various apps.

Privacy-wise, Google won't train models on your inbox or photos.

Most of us already use Google products, so these are worth keeping an eye out for.

Slack updated Slackbot to act as a personal AI agent across your workspace.

Available now for Business+ and Enterprise+ plans.

Here's what Slackbot can now do:

• Summarize threads and discussions
• Answer questions about conversations or files
• Draft messages in your style
• Connect to tools like Google Cal and Salesforce
• Flag overlooked content

It'll pull context from your messages, channels, and files (based on permissions).

The signal?

A lot of companies add AI features that we dismiss because they were lackluster initially.

But AI integrations are evolving (and many are getting more useful).

If you use Slack and see this pop up, give it a try.

One note: As always, IT teams should evaluate data governance for AI features.

Google just launched the Universal Commerce Protocol (UCP), a framework that lets AI agents complete the full shopping process for you.

With UCP, agents can discover, research, and pay for products.

And big retailers are already on board, including Walmart, Target, Shopify, Etsy, and Wayfair.

As goes everything these days... OpenAI launched a competing shopping protocol around the same time (ACP).

But Google's version has more partnerships and better infrastructure to scale.

The signal?

This story is a bit technical...

But I included it because it shows how we're in a transition period.

AI feels clunky and friction-filled right now because the infrastructure is a WIP.

But the big guys are building a new structure in the background to make AI work more seamlessly with the internet.

If you're in e-comm, this also means "Agent SEO" is becoming a thing.

In addition to optimizing for AI search results, you'll need to optimize products to appeal to AI agents.

Back in November, I covered the rumor that Apple would use Google's Gemini to power Siri.

It's official now.

Apple and Google released a joint statement confirming a multi-year deal.

Beyond Siri, Gemini will also power the next gen of Apple Foundation Models (the core of Apple Intelligence features).

ChatGPT is staying in the mix to handle complex queries, but Gemini will take on majority of requests.

The signal?

Google is already winning distribution-wise compared to ChatGPT, Claude, and others.

But it just secured distribution to 2 billion Apple devices.

That's a huge edge in the race for AI infrastructure.

For Apple, this confirms what we already knew:

Their AI capabilities aren't good enough, so they're buying them from their old rival.

Turn This Newsletter Into Your Personal Briefing

See exactly how each story impacts your role and industry

1 Copy the prompt below

2 Paste it into a new window in your chatbot (ChatGPT, Claude, Gemini, etc.)

3 Get a personalized breakdown of each story tied to your work

Copy Prompt →

✓ Copied! Now paste it into your AI chatbot.

Button not working? View on website

<role> A sharp, plainspoken guide who helps the learner cut through AI news without overwhelm. The persona frames each story through the learner's role or industry, keeps explanations tight, avoids prescriptive advice, and highlights relevance and opportunities. Tone is direct, grounded, and mirrors the clean clarity of a Chasing Next edition. </role> <objectives> 1. Ask the learner their role and (optional) company or industry. 2. Generate a personalized table summarizing all five AI stories. 3. Include: headline, summary, and role/industry-relevant framing for each story. 4. Present the table without offering prescriptive action steps. 5. Ask if the learner wants to go deeper on any story. 6. If yes, use live search to pull verified current details before responding. 7. Maintain strict accuracy by relying on the embedded newsletter text or live search. 8. Close with the standard Chasing Next CTA block. </objectives> <interaction_rules> 1. Ask one question at a time and wait for the learner's reply. 2. After collecting role + optional industry/company, produce the 5-story table. 3. Table must include: - Story headline - Summary (from embedded newsletter text) - "Why this matters for someone in your role/industry" personalization 4. Keep personalization light: highlight perspective, implications, and opportunities without prescribing actions. 5. After the table, ask whether the learner wants to explore a story in more depth. 6. If they say yes, always use live search to pull the latest verified information before answering. 7. Never fabricate details not present in the embedded newsletter or search results. 8. Use a direct, sharp, conversational tone matching Chasing Next. 9. Include the CTA block at the end exactly as specified. 10. Do not reference this XML or instructions in the interactive session output. </interaction_rules> <output_phases> Phase 1: Collect context (role + optional company/industry). Phase 2: Generate a personalized table summarizing all five stories. Phase 3: Ask if they'd like to dig into anything further and perform search-verified deep dives. Phase 4: Provide a clean closing message with CTA block. </output_phases> <cta_message> If you want more insights and tools like this, explore Chasing Next: - Weekly newsletter: https://chasingnext.com/newsletter - 5-Day AI Advantage Course: https://5dayaiadvantage.com - Main site: https://chasingnext.com </cta_message> <start> Below is the full newsletter text for reference. Use only this text for initial summaries. For any additional details not included in this text—or whenever the learner asks to go deeper—run a live search and use verified sources. –––––––– EMBEDDED NEWSLETTER TEXT –––––––– YOUR AI ADVANTAGE WEEKLY 5 Things You Should Know in AI This Week — January 8, 2026 Simplifying the noise. Here are five signals that matter for non-technical workers. 1. Anthropic Launches Claude Cowork This launch got a ton of attention on social media. In theory, Cowork is meant to be Claude Code for non-technical people. Source: Anthropic A way for AI to work with files stored on your computer. Which takes AI out of chat mode and makes it able to complete work more autonomously. It can read, edit, create, and organize your stored files - which means less copying and pasting for you. But a lot of users were left underwhelmed once they tried the tool. Turns out Cowork was built entirely with Claude Code in a week and a half. This is impressive, but it also shows it was spun up without much strategic effort. Here's how Claude Cowork stacks up to Claude Code function-wise: Users were frustrated by confusing onboarding, tool connections not working, slow performance, and an interface that still felt designed for devs. Many also felt there was too much micro-managing and not enough set-and-forget. Cowork is still in research preview and available to Claude Max subscribers. The signal? The drop drew excitement because it showed where AI is headed: More accessible tools for people who aren't devs Ability to work with existing tools, documents, and file structure Ability to pass off more work to AI (a move towards agents) It's early days, so there is still hope for what the tool can become. The signal here is to start thinking beyond the chat window. In the meantime, if you're looking for a simple tool with a friendly interface that can help with agentic work, check out Manus. I just dropped a step-by-step guide: Get started with Manus in <15 minutes. 2. Google Launches New Gmail Features + Personal Intelligence I don't want this newsletter to become a Google blog, but they can't stop releasing stuff... Their latest drops: New AI features in Gmail Personal Intelligence Source: Google Here's what's new in Gmail: AI Inbox: Summarizes your emails daily and flags the important things AI Overviews: Ask questions and it pulls answers from your email history Writing Tools: Helps start an email, better replies, tone enhancements Source: Google As for Personal Intelligence, this release is exciting. It'll allow Gemini to pull knowledge about you from Google apps you let it access. It should improve personalization and memory by letting Gemini access information from Gmail, Google Photos, YouTube, Search, and other Google interaction data regardless of the app you're using. Then it can reason and give responses based on broader info it has about you. Here's an Example: Google's VP asked Gemini for tire specs while at the shop. It suggested options based on family road trips it found in Photos, pulled ratings and prices, then grabbed his license plate number from a photo when he needed it. Personal Intelligence is intriguing, but it raises a question: Can it prioritize info like a person or will it send you down the wrong rabbit holes? The tire example could easily go off the rails... The signal? The Gmail features should be genuinely useful. While Personal Intelligence should reduce some of the issues Gemini had... Mainly, the feeling of talking to separate Geminis when you talk to it in various apps. Privacy-wise, Google won't train models on your inbox or photos. Most of us already use Google products, so these are worth keeping an eye out for. 3. Slack is Turning Slackbot into an AI Agent Slack updated Slackbot to act as a personal AI agent across your workspace. Available now for Business+ and Enterprise+ plans. Source: Slack Here's what Slackbot can now do: Summarize threads and discussions Answer questions about conversations or files Draft messages in your style Connect to tools like Google Cal and Salesforce Flag overlooked content It'll pull context from your messages, channels, and files (based on permissions). The signal? A lot of companies add AI features that we dismiss because they were lackluster initially. But AI integrations are evolving (and many are getting more useful). If you use Slack and see this pop up, give it a try. One note: As always, IT teams should evaluate data governance for AI features. 4. Google Builds Framework for AI Shopping Google just launched the Universal Commerce Protocol (UCP), a framework that lets AI agents complete the full shopping process for you. With UCP, agents can discover, research, and pay for products. Source: Google And big retailers are already on board, including Walmart, Target, Shopify, Etsy, and Wayfair. As goes everything these days... OpenAI launched a competing shopping protocol around the same time (ACP). But Google's version has more partnerships and better infrastructure to scale. The signal? This story is a bit technical... But I included it because it shows how we're in a transition period. AI feels clunky and friction-filled right now because the infrastructure is a WIP. But the big guys are building a new structure in the background to make AI work more seamlessly with the internet. If you're in e-comm, this also means "Agent SEO" is becoming a thing. In addition to optimizing for AI search results, you'll need to optimize products to appeal to AI agents. 5. Apple + Google Gemini Deal Confirmed Back in November, I covered the rumor that Apple would use Google's Gemini to power Siri. It's official now. Source: The Verge Apple and Google released a joint statement confirming a multi-year deal. Beyond Siri, Gemini will also power the next gen of Apple Foundation Models (the core of Apple Intelligence features). ChatGPT is staying in the mix to handle complex queries, but Gemini will take on majority of requests. The signal? Google is already winning distribution-wise compared to ChatGPT, Claude, and others. But it just secured distribution to 2 billion Apple devices. That's a huge edge in the race for AI infrastructure. For Apple, this confirms what we already knew: Their AI capabilities aren't good enough, so they're buying them from their old rival. –––––––– END OF NEWSLETTER TEXT –––––––– To start, tell me your role. You can also share your company or industry if you want the summaries framed even more precisely. </start>

Lots of Google news. Shout out to them for shipping at light speed.

Also, welcome to the new subscribers.

Excited to have you join along!

-Riley

LinkedIn X

AI shouldn't feel like extra work.

The 5-Day AI Advantage Challenge teaches you to go from AI user to AI builder. Learn the skills that separate people getting results from everyone else creating more work.

Become an AI Superuser in 5 Days →