- Charbel X
- Posts
- Gemini 2.0, Novice-User-Friendly Design, Lessons From a Startup Failure
Gemini 2.0, Novice-User-Friendly Design, Lessons From a Startup Failure
Google's catching up with the AI rush.
Charbel Zeaiter
December 13, 2024
Gosh, so much has happened overnight, my brain is fried! I’m yet to experience Google’s Gemini 2.0 - but it’s the current talk of the town in AI Communities.
As always, so much in today’s newsletter.
Wishing you a superb weekend.
Yours in Wonder,
Charbel
Founder of Velvet Onion, Faster Zebra and more to come …
Today’s Highlights
AI: Google’s Gemini Grew Up: Gemini 2.0 with Incredible AI Agent Capabilities
Design: Design Like A Savvy Designer, Critique Like A Newbie User: On User-Friendly Designing
Science & Tech: Present and Future of Self-Driving Cars
Founding: Why did DirectAI Shut Down Despite Having Good Innovative Bits to Offer?: Lessons from a Startup Failure
Product: Building Hypnotic Product Pages with AI
Today’s AI image: Robotaxis: Turning Sci-fi Into Reality
Quote for the day: Another Reason to Live In the Present
AI
Google’s Gemini Grew Up: Gemini 2.0 with Incredible AI Agent Capabilities
Google’s just dropped Gemini 2.0, and it’s not a tad less than its competitors.
With its boosted multimodal smarts, seamless tool integration, powerful research capabilities, and a handful of experimental projects, it’s clear the tech giant’s pushing the boundaries of AI agents.
Gemini 2.0
Say hello to Gemini 2.0 Flash, the speedier, more efficient model that leaves the hefty 1.5 Pro in the dust on several benchmarks, all while keeping pace on speed.
This new version can handle images and multilingual audio straight off the bat, while processing text, code, images, and video with ease.
On the freebie front, there’s Gemini 2.0 Stream Realtime. Unlike ChatGPT Pro’s $200/month price tag, this one’s free and lets you interact through text, voice, video, or even screen-sharing.
Deep Research:
Say hello to Deep Research, an AI-powered assistant designed for heavy-duty research.
Now available in Gemini Advanced ($20/month) on both desktop and mobile, this tool can create multi-step research plans, scour the web for info, and whip up detailed reports with handy source links.
Why is this a big deal?
OpenAI may have had the holiday hype, but Google’s come in hot with Gemini 2.0, cranking up the AI game.
With jaw-dropping upgrades, this could be one of the biggest steps we’ve seen towards consumer-ready, agentic AI. Projects like Astra could also set a new bar for AI interactions as we roll into 2025. Time to buckle up.
Also in AI
Design
Design Like A Savvy Designer, Critique Like A Newbie User: On User-Friendly Designing
As a designer, jumping into a project early and experiencing it yourself is vital. You need firsthand insight, not just secondhand knowledge, to truly grasp user struggles and design with clarity.
The “Minion” Moment
A hilarious yet crucial lesson: relying on jargon that your team understands doesn’t mean the end user will.
Design terms might seem intuitive to experts but can confuse newcomers—always question things, no matter how trivial it seems.
Expertise Blindspots
Despite decades of collective experience, even seasoned designers can overestimate what a user knows.
Expertise doesn’t necessarily translate into empathy for a novice user’s perspective, which is why direct involvement is non-negotiable.
Heuristic Evaluation
Heuristic evaluation is a quick, cost-effective way to spot major UX issues without the need for formal reports.
It's about identifying glaring problems based on established usability principles, and it helps designers stay grounded in real-world user needs.
Avoid Becoming a Perpetual Intermediate
Once you’re familiar with a project, you risk becoming too comfortable and losing that ‘novice’ perspective.
Regularly revisit the struggle of being a beginner to uncover new insights that will refine your design approach.
By experiencing your designs firsthand, questioning assumptions, and taking up a mix of evaluation methods, you’ll steer clear of common pitfalls and create products that hit the mark.
Also in Design
Science & Tech
Present and Future of Self-Driving Cars
Once sceptical about the hype surrounding autonomous vehicles, it’s time to rethink. What was once seen as a distant fantasy is now becoming a reality, particularly with robotaxis leading the charge.
The Rise of Robotaxis
Cities like Wuhan and San Francisco are at the forefront of adopting self-driving taxis.
In Wuhan, 3 out of every 100 taxis are robotaxis, and in San Francisco, Waymo has seen its ride numbers soar, surpassing 300,000 monthly rides.
As robotaxis take off, they're quickly chipping away at the market share previously dominated by traditional ride-hailing services.
Self-Driving Beyond the Tech Bubble
Historically, tech products that appeal to early adopters often fail to reach the masses, but robotaxis are proving to be the exception.
Waymo is expanding its operations into multiple U.S. cities, and Tesla’s rising popularity is helping ease the public into accepting self-driving technology.
Trust is Building
Public trust is a significant hurdle, but things are improving. Cities with robotaxis, like Phoenix and San Francisco, are reporting a higher confidence level in autonomous vehicles than the average American.
As people interact more with the technology, their comfort levels increase.
Safety? *Silence*
While accidents, like a pedestrian being hit by a Cruise vehicle in San Francisco, have highlighted the risks, the safety track record of self-driving cars is improving.
The Road Ahead
While still a work in progress, the momentum behind autonomous vehicles is undeniable.
With growing acceptance, improving safety, and impressive technological strides, self-driving cars are starting to shift from science fiction to everyday transportation.
Also in Science & Tech
Founding
Why did DirectAI Shut Down Despite Having Good Innovative Bits to Offer?: Lessons from a Startup Failure
A startup with seemingly perfect product-market fit and extraordinary innovations to sell had to wind up. Here’s
Why?
The use cases offered by DirectAI were too specific that scaling touched the ceiling way early.
When you offer a solution to a problem faced by no more than 5 people, you can scale your market share merely up to 5 people.
The only way to expand revenue in such cases is by innovating yet another product for the same 5 people. But it also has its limits and it’s not apparent that they will need your new creation as much.
Besides, DirectAI struggled to provide its customers custom solutions with venture-scale efficiency.
3 Lessons To Learn (and Note)
#1 Develop a Witty Sense of Bad Ideas and Potential Not-Gonna-Work-s
A smart startup learns from its mistakes. A smarter startup learns to avoid mistakes.
Almost all the faults and failures can be dodged with a bit of expert foresight and calculative decision-making.
For instance, conduct sales calls instead of user interviews. Because in user interviews, users attempt to justify the deal they’ve already struck and hence, might give you false positives.
Consider sales calls on the contrary: all the talk prospects do straightly imply what they think about your product and offer.
You can judge their true intent clearly on such calls. How does it help? Helps you know whether to adopt or to reject the idea right away.
#2 “Our Product Helps You Solve Your Problem” < “Our Product is Your Solution”
Customers don’t want tools. They want complete solutions to their problems. For example, a developer would like a code debugger rather than merely a bug detector.
DirectAI's evolution from providing classifiers to delivering integrated solutions (e.g., video provider plugins) highlighted the importance of meeting customers where they operate.
#3 Sometimes Collabs Don’t Monetise
Partnerships did provide exposure to DirectAI but barely translated to sales.
Turned out direct sales and marketing were more effective for gaining customers.
Product
Building Hypnotic Product Pages with AI
You understand your product and AI understands your data. A collab would be able to grow the best ecommerce product pages for your product. Here’s
How to make it happen?
Borrow an AI hand for Speedy Generation of Best Quality Pages: AI can create detailed and convincing product descriptions quickly, saving you time, cost and a load of trouble, particularly if your business has a large inventory.
A/B Test Numerous Variations of Pages the Least-Effort Way: AI is like an ever tireless brain that can easily carry out the creation and testing of multiple variations of product pages.
Besides, identifying the most effective layouts, headlines, and descriptions, all data-driven is no longer an impossible task.
Let AI take a look at your customer data and personalise their purchase experience: AI can tailor product pages to individual customers Like it could suggest a customer complementary items based on their past purchases.
Today’s AI Image
Robotaxis: Turning Sci-fi Into Reality
Quote of the Day
Another Reason To Live in the Present
"The present is the only thing that has no end.."
Erwin Schrödinger
What we’re working on
 Velvet Onion & Friends The new Velvet Onion & Friends will be launched soon. It’s our latest evolution, helping companies build products. It’s more than services. |  Faster Zebra February 2025 - the product and venture school journey begins. Whitepaper launching in January. |
