Skip to main content

Posts

Google now lets you share custom Gemini AI assistants called Gems

Google’s Gem Sharing for Gemini AI Google now lets you share your custom Gemini AI assistants, which the company calls Gems.  What are Gems? Gems are AI assistants made for specific tasks.  They started off as part of the Gemini Advanced paid plan.  You could use them for things like learning coaching, writing help, brainstorming, career advice or coding.  What’s new Now you can share Gems with others. Family, friends, or coworkers.  Sharing works like sharing a file in Google Drive.  You control who can view, use, or edit the Gem.  Why this matters It makes Gems easier to use by more people, not just paid users.  It stops duplication. If many people need a similar tool, one shared Gem is better than many similar ones.  It helps with teamwork: planning trips, writing together, or designing meals.  Rollout details Initially only for Gemini Advanced, Business, and Enterprise users. Over 150 countries.  Later G...
Recent posts

Meta Shows Glasses That Could End the Smartphone Era

Meta wants us to trade in smartphones for smart glasses.  Zuckerberg spoke at Meta Connect 2025. He said phones reduce the sense of presence among people.  The product: Meta Ray-Ban Display. It has cameras, speakers, microphones, and an AI helper. The display is placed so it won’t block your view. It shows apps like Instagram, WhatsApp, Facebook, plus maps and live translation.  Meta also showed a wristband. The band picks up signals from your hand and brain when you make a gesture. You can write text without speaking. Just move your fingers like writing.  Zuckerberg said he could type about 30 words per minute with that setup. Research shows people type about 36 words per minute on a phone. Average users with Meta’s system get around 21 words per minute.  Meta has lost a lot of money since 2020 in its Reality Labs division. But this project shows what some of that money has been going into.  Meta hopes these glasses will compete with Appl...

ChatGPT Explained: Uses, Updates, and Issues

What ChatGPT is ChatGPT is an AI chatbot made by OpenAI .  It produces text when you give it prompts.  It uses large language models (LLMs) such as GPT-4 and more recent ones.  Growth & scale Released in November 2022.  Thousands of updates since. Many new features and improvements.  ~300 million weekly active users as of mid-2025.  Recent updates & features ChatGPT can now work in different “modes” like “Auto”, “Fast”, “Thinking”.  New tools for coding, calendar handling, research, shopping, etc.  Better protections for minors. Stricter rules around sensitive content.  A cheaper plan in India (ChatGPT Go) launched for about ₹399 per month.  Challenges & criticisms Legal issues: lawsuits over copyright, concerns about defamation from false statements.  Risk of misuse or harm in sensitive conversations (mental health, privacy, etc.).  Reliability issues: sometimes makes mistakes or “hallucinates” (i.e. g...

Google adds Gemini to Chrome in the U.S. with new smart browsing tools

What Google’s Doing 1. Gemini in Chrome is wider now It used to be only for Google AI Pro and Ultra users. Now any Mac or Windows desktop user in the U.S. with Chrome can use it.  2. It helps with reading hard stuff If a page is confusing, you can click the Gemini icon in Chrome. Ask Gemini to explain or simplify what you see. Like making a recipe gluten-free.  3. Work with many tabs at once Gemini now can pull info from several tabs. Want to compare hotel vs flight vs vacation spots? Gemini can help.  4. Remember pages you saw before You’ll be able to ask Gemini: “Where did I see that walnut desk last week?” or: “What was the blog about back-to-school shopping?”  5. Tighter with other Google stuff Gemini will link with Calendar, Maps, YouTube. So you can get schedules, maps or video moments without leaving the page.  6. Can do tasks for you Gemini will help with small chores: booking haircuts, ordering groceries. It’ll add things to your cart. You ...

Open AI’s Next Steps with Thomas Wolf

Thomas Wolf, co-founder and chief science officer at TechCrunch Hugging Face, spoke at Disrupt 2025. He focused on making AI open, shared, and fair.  He began by asking who builds AI, who shares it, and who scales it. He said that AI should not be only inside closed labs or held by big firms.  Wolf believes open source and global cooperation will shape AI’s future. He thinks moonshot experiments and shared research play large roles.  He noted that open tools and libraries let many people try ideas. He pointed to his work at Hugging Face.  Wolf co-founded Hugging Face. He helped build the Transformers and Datasets libraries. He led the BigScience Workshop. That work produced BLOOM, a large language model.  He also wrote resources for how folks learn AI and build it. Among them are Natural Language Processing with Transformers and The Ultra-Scale Playbook.  Wolf argued for openness in both research and practice. He wants models that anyone can us...

Nvidia invests $5 billion in Intel, plans AI chip work

What happened Nvidia is buying $5 billion worth of Intel common stock.  Price per share is $23.28.  This makes Nvidia one of Intel’s larger shareholders (roughly 4-5%).  What Intel and Nvidia will work on together They plan to co-develop chips for data centers and personal computers.  Intel will design custom x86 CPUs for Nvidia’s AI infrastructure.  Intel will also build system-on-chips (SoCs) that include Nvidia’s RTX GPU chiplets for PCs.  Why this matters Intel's stock jumped ~25-30% after this news.  It gives Intel a strong boost. The company has struggled recently.  For Nvidia, this move brings its GPU tech even closer to CPUs. That may help improve performance in AI setups.  Potential risks or unknowns It’s not certain when these new Intel-Nvidia chips will reach the market .  Integration of CPU and GPU work, especially combining them in consumer-level chips, is technically tough. Regulatory approval might be required...

Google and PayPal join forces on agentic commerce

What’s happening PayPal and Google signed a multi-year deal. Google will use its AI tools to build new shopping features. PayPal will plug its payment, identity, and personalization tools into Google’s products. Key parts of the deal They won’t say yet what exactly “agentic shopping” means in practice. Google gives AI tech; PayPal gives its global payments system, identity and personalization. The two also back Google’s new Agent Payments Protocol . Over 60 merchants and financial groups already support the protocol. Integration & use PayPal will be a main payment option in Google Ads, Google Cloud, and Google Play. Services like PayPal’s branded checkout, Hyperwallet payouts, and PayPal Payouts will be integrated. What this could mean Shoppers might be able to buy things through AI agents using PayPal. Google’s AI could better anticipate what users want to buy. Payments and identity verification may become smoother, since PayPal already hand...