😏Woke Barbie😏

Mattel has released its first-ever Barbie with autism — and for once, this isn’t a vague “inclusive” gesture doing the bare minimum. This doll was co-designed with autistic advocates, and it shows 💜

She comes with a working fidget spinner, noise-cancelling headphones, and an AAC (augmentative and alternative communication) device. Her gaze is slightly averted, her outfit is intentionally loose and sensory-friendly, and nothing about her design screams “educational toy”. She’s just… Barbie. With autism. And that distinction matters.

For families like Precious Hill’s — a mum and daughter both diagnosed autistic — the impact was immediate. Her five-year-old daughter recognised herself straight away. Same tools. Same needs. Same normality. That moment of recognition? That’s the bit policy papers and awareness campaigns rarely deliver ✨

What’s especially significant is that this is a girl Barbie on the spectrum. Autism in girls is still routinely missed, misdiagnosed, or discovered much later in life — often when they become mothers themselves. Seeing a female, non-verbal-inclusive representation in one of the world’s most recognisable toys quietly challenges decades of stereotypes about who autism “looks like” 👧🏽

This isn’t claiming to represent the entire spectrum — and no one serious expects one doll to do that. But it does reflect something important: the tools that support independence, the reality of sensory overload, and the fact that disability doesn’t cancel joy.

And yes, it’s still Barbie. Which is kind of the point.

Toys shape how children understand themselves before society has a chance to tell them they’re “too much”, “too quiet”, or “different in the wrong way”. A mainstream brand choosing to reflect autistic experiences — without pity, pathology, or performative messaging — helps normalise difference early. That’s not nothing. That’s culture shift, just scaled for a toy shelf 🧠💬

Representation doesn’t fix systems. But it does tell kids — and adults diagnosed far too late — that they’re not invisible.

You can read more here 

https://edition.cnn.com/2026/01/16/health/autism-barbie-spectrum

🤖 AI gets personal🤖

Google has quietly launched Personal Intelligence, a new beta feature that lets Google’s Gemini assistant reason across your Gmail, Photos, YouTube and Search history — without you having to specify where the info lives. One question, multiple apps, stitched together into a single answer.

On the surface, this looks like another “AI gets more personal” update. Underneath? This is Google activating an advantage no rival can realistically replicate.

• Gemini can reason across apps, not just fetch isolated facts. Emails, photos, videos and searches are combined into one response.

• It works across text, images and video, pulling specifics like licence plates from Photos or context from old Gmail threads.

• It’s off by default, opt-in, and users choose exactly which apps to connect.

• Google says your inboxes and photo libraries are not used to train models directly — they’re referenced, not absorbed.

• Rolling out first to Gemini AI Pro and Ultra subscribers in the US, with plans to expand later.

So every AI company is talking about “memory”:

• ChatGPT remembers preferences.

• Claude remembers conversations.

• Perplexity remembers searches inside its own app.

Google remembers your digital life.

Years of emails. Every photo you’ve taken. Your YouTube habits. Your search history stretching back to the mid-2000s. That data already exists, already structured, already linked to identity — and now it’s being wired directly into an AI assistant.

This is the AI moat nobody else has:

• OpenAI can’t read your Gmail.

• Anthropic can’t see your Google Photos.

• No startup can recreate 20 years of behavioural data at global scale.

As frontier models converge in quality, personal context becomes the differentiator — and Google owns the pipes.

Google is doing the right things on paper: opt-in, transparency, explanations of where answers come from, and the ability to turn personalization off per chat. But let’s be honest: this still asks users to trust one company with unprecedented insight into their lives — now interpreted, summarised and acted on by an AI layer.

So what?

This isn’t about Gemini getting smarter. It’s about Google quietly shifting the AI race away from benchmarks and towards data gravity. The models are catching up to each other. The question now is: who already knows you best?

For everyone else in AI, the problem is brutal and simple:
How do you compete on personalization when your competitor has the user’s entire digital history — and you’re starting from a blank chat box?

📎 Read more:
Gemini introduces Personal Intelligence – Google
https://blog.google/innovation-and-ai/products/gemini-app/personal-intelligence/

😬 Grok😬

After days of backlash, government pressure and very public harm, X has finally said it will stop Grok from “undressing” images of real people in jurisdictions where it’s illegal. Ministers called it progress. Ofcom called it “welcome”.😑

Campaigners called it too late.

They’re right.

Because this wasn’t a technical glitch. It was a design choice.

Grok didn’t accidentally stumble into generating sexualised images of real women. It shipped with the capability, stayed live while abuse spread, and only changed course once regulators, victims and the prime minister were all shouting at once. Even then, reporting from the Guardian shows the supposed crackdown is porous at best — with a standalone Grok tool still happily producing sexualised videos of real women that can be posted straight back onto X.😤

More worryingly perhaps it wasn't a decision because that is wrong" whether its illegal or not, it is just a mere compliance measure.😖

What’s striking is how reactive everything has been. Safeguards only arrived after harm. Geoblocking only arrived after regulators intervened. And enforcement still appears optional, inconsistent, and trivially easy to bypass with basic tools like VPNs.📲

So what?

This saga exposes a deeper truth about the AI industry:

The sector is still optimising for growth, engagement and shock value over well anything else. Other model operators which appear to have less deranged people running them recognise where the winds of public opinion may lie but it is not consistent clearly 

Read more:

BBC – X to stop Grok AI from undressing images of real people after backlash
https://www.bbc.co.uk/news/articles/ce8gz8g2qnlo

The Guardian – X still allowing users to post sexualised images generated by Grok AI tool
https://www.theguardian.com/technology/2026/jan/16/x-still-allowing-sexualised-images-grok-ai-nudification

—

📶Google appeals📶

😮‍💨 In the least surprising news of the year so far… Google is appealing its monopoly ruling

Google has officially appealed the US court decision that ruled it illegally held a monopoly in online search — a move so predictable it barely qualifies as breaking news. The company says the original ruling “ignored reality”, arguing people use Google because they want to, not because the entire internet quietly nudges them there by default 🔍

Sure.

The appeal targets last year’s landmark judgment by Judge Amit Mehta, which stopped short of breaking Google up but did order a set of remedies designed to loosen its grip — including forcing Google to share parts of its search index and allow rivals to display Google results as their own. Google is now asking for those fixes to be paused, claiming they threaten privacy, discourage innovation, and unfairly punish success 🚨

The irony is thick. Google’s defence rests on the idea that competition is fierce and innovation is constant — while simultaneously fighting tooth and nail to avoid sharing the very infrastructure that made it untouchable in the first place. If competition is so intense, why the panic over letting others see under the hood? 🤔

Judge Mehta explicitly noted that AI has changed the trajectory of search, which is precisely what makes this appeal so telling. As generative AI reshapes how people find information, Google is desperate to defend its default position — especially as regulators in both the US and EU circle its AI-powered search summaries and data usage practices 🤖

And let’s not forget the backdrop: Google’s parent company Alphabet just hit a $4 trillion market cap.

The idea that this is a fragile innovator being unfairly constrained is… ambitious storytelling 💸

🧠So what?

This isn’t Google denying that the market is changing — it’s Google arguing for how that change should be managed. The appeal buys time and clarity as courts, regulators and companies all grapple with what competition looks like when search, AI and distribution blur together. The outcome will matter less for today’s search box, and far more for who gets to shape the next interface layer of the internet ✨

Read more:

BBC News – Google appeals landmark antitrust verdict over search monopoly
https://www.bbc.co.uk/news/articles/clyn0ek5rdpo

🏥Chat GPT health🏥

OpenAI has spent the week doing two very loud things at once: buying a healthcare start-up for over $100m and quietly trialling ads inside ChatGPT. Separately, both are “interesting”. Together, they’re… revealing.

First, the deal. OpenAI has acquired Torch, a US health-tech start-up founded in 2025 that promises to unify your scattered medical records — labs, hospitals, wearables, consumer tests — into what its founders call a “unified medical memory for AI”. That tech is being folded directly into ChatGPT Health, a new feature that lets users upload medical records and talk through test results, diets, workouts and appointment prep with a chatbot 🤖

OpenAI insists this data will be private, siloed, not used to train its models, and definitely not used for diagnosis or treatment. Campaigners say: fine, but health data is the most sensitive data most people will ever share, and “trust us” is not a regulatory framework 🧬

Now for the second move: ads. ChatGPT will soon start showing sponsored banners after prompts for some users — starting in the US — as OpenAI looks for ways to stop burning billions of investor cash. Ask about holidays, get holiday ads. Ask about restaurants, get restaurant ads. Totally normal internet behaviour resumes💸

But here’s where the irony hits.

OpenAI is actively encouraging users to upload medical records, fitness data, nutrition logs and health histories — while at the same time building an advertising business. The company swears there’s a firewall between health data and ads. Analysts, privacy experts and anyone who’s ever used the internet before are… sceptical 🧱

Because once your chatbot knows your cholesterol, your BMI, your diet habits and your exercise gaps, the dystopian joke basically writes itself:
“Your blood sugar looks high. Have you tried lifestyle changes?”
Sponsored: Ozempic.

No one is accusing OpenAI of doing this yet. But the structural incentive is obvious. Health is one of the most lucrative advertising categories on earth. And as one expert put it, AI firms are “leaning hard” into personalisation to boost value. Trusted medical adviser + ads is not a neutral combination 🧠

This is also happening in a market where OpenAI reportedly lost $8bn in six months, only 5% of its 800m users pay, and the AI bubble is increasingly being whispered about as, well… a bubble. Ads aren’t a philosophical pivot. They’re a revenue survival strategy 🫧

🧠So what?

This isn’t just about OpenAI. It’s about the collision of health, trust and monetisation in consumer AI. When platforms position themselves as personal, intimate, always-on advisers — and experiment with ads — the line between “supporting wellbeing” and “influencing behaviour” gets dangerously thin.

If AI is going to sit between patients and their health decisions, the safeguards need to be stronger than blog posts and promises. Because once the model knows your body, your habits and your anxieties, the question isn’t whether it could be used to sell to you — it’s whether anyone will be able to prove when it isn’t 🚨

Read more:

Silicon Republic – OpenAI acquires Torch for more than $100m in healthcare push
https://www.siliconrepublic.com/business/openai-torch-acquisition-100m-health-tech-healthcare-medicine-ai

BBC – ChatGPT to trial adverts as OpenAI looks for new revenue
https://www.bbc.com/news/articles/cvgjn012k3do

BBC – ChatGPT Health raises privacy concerns as users share medical data
https://www.bbc.co.uk/news/articles/cpqy29d0yjgo

📈 The tools behind the tech📉

📦Product📦

• How to build a platform?
• The Long Now of the Web: Inside the Internet Archive’s Fight Against Forgetting
  

📏Design📏 

• Using AI as a Design Engineer

👩🏿‍💻Code👩🏿‍💻

• A LLM trained only on data from certain time periods to reduce modern bias
• The Next Two Years of Software Engineering

🏢The business behind the tech🏢

•  FTSE 100 CEOs earn more than average worker’s yearly pay by noon on 6 January
• The SaaS Selloff: AI and Interest Rates
  
  

🙌Amplify with Bloomberg🙌

Discover Innovative and Exciting Insights from Bloomberg’s Data and Engineering Teams  

On Feb 25th, Bloomberg is joining forces with Colorintech for another installment Amplify, an event dedicated to amplifying diverse voices in the world of tech! 

The evening will feature a series of dynamic lightning talks and a special panel discussion with leading voices at Bloomberg showcasing innovation and insights from Bloomberg’s Data and Engineering teams.

Additionally, the team are expecting to have roles in the following areas:
Software Engineers: 4+ years experience with proficiency in a high level language like C++ or Python and interest in systems design;

Data Management professionals: (including Data Engineers and Data Quality) with 4+ years experience 

Technical Account Manager: 3+ years experience. 

So if you’re also interested in speaking to someone about your next career step at Bloomberg, I’d strongly advise registering your interest. 

Check out the key details below: 

Date: Wednesday 25th Feb 2026 

Time: 17:30 - 20:30 UK Time 

Where: Central London 

As this event has limited spaces, you’ll need to register your interest to attend using the form below: 

Application Link: https://bloomberg.avature.net/su/0cab465ddb1412ae 

Application Deadline: 18th Feb 2026 

After applying, the Bloomberg team will reach out to you and provide you with next steps if they have invited you along. 

This is a great opportunity for anyone who wants to learn, be inspired, and build connections with Bloomberg Professionals who are shaping the future of technology!

👉EIR program

Partnerships can make or break a startup, but founders are rarely taught how to 'build relationships'... until now.

Whether you're a B2B, B2C or B2B2C, understanding how to start, build and scale thriving partnerships with corporate organisation and investors can be a huge factor in your startups success - notable examples being:
General Motors partnering with Lyft (leading ride-sharing startup);
Coca-Cola partnering with Hivery (AI Startup focused on optimizing shelf space and vending machine operations);
Anytime Fitness partnering with DeltaTrainer (Digital personal training platform).

With this session of Entrepreneur in Residence, we're going to share tips, tricks and lessons for creating successful partnerships for your business.

What to expect:
A sparknotes workshop on navigating corporate partnerships and understanding corporate jargon;
An AMA with a guest speaker (to be revealed soon );
Networking time.

This opportunity is designed for founders and builders at any stage thinking about building corporate and investor partnerships!

You don’t need to be full-time or funded. If you’re serious about building, you’re welcome.

Please see the key details below:

Date: Tue 3rd Feb 2026
Time: 18:00 - 20:30 BST
Location: London, UK!
Event Page: https://luma.com/EIR32026

 Spots are limited – register now to secure your place 

😃What we are consuming😃

💸Why VCs are betting big on AI security

💰Sequoia to invest in Anthropic, breaking VC taboo on backing rivals

🚀MIT's 10 Breakthrough Technologies

☀️Sun simulator

📲Tik tok for learning