- Charbel X
- Posts
- Amazon's New AI Dynasty Begins, Great Brands Ditch Hypes, Pre-Mortem Meetings and more
Amazon's New AI Dynasty Begins, Great Brands Ditch Hypes, Pre-Mortem Meetings and more
Trends rise and fall. So do you, if you keep catching up with them.
Amazon has invested heavily in Anthropic’s Claude, and they’ve just released a whole swather of AI tools. Let’s brace ourselves for the onslaught of AI generated ‘creative’ content.
A few years ago, I posed the question, “when will Artificial Intelligence” simply be referred to as ‘intelligence’? Yikes.
Yours in Wonder,
Charbel
Founder of Velvet Onion, Faster Zebra and more to come …
Today’s Highlights
AI: Amazon’s Nova AI Model Family: A Top-Notch Team Text, Image and Video Generation Models
Design: Running with Hypes and Trends is Not Always a Good Trade-Off
Science & Tech: Brain Mapping Using Viruses: Creating Neuron "Barcodes"
Founding: Product-Market Fit isn’t a one-time task: It Needs To Be Maintained Or Your Numbers Die
Product: Pre-mortem Meetings Used by Meta and Instagram: A Pre-mortem of Product Before Launch
Today’s AI image: Amazon’s AI Family
Quote for the day: A Vital Life Principle from Orwell
AI
Amazon’s Nova AI Model Family: A Top-Notch Team Text, Image and Video Generation Models
Amazon has thrown its hat into the generative AI ring with Nova, AI models designed to handle text, images, and video; A whole family of them.
A signal towards its most ambitious foray yet into consumer-focused generative AI.
All You Need To Know
Nova’s roster includes four text-specific models—Micro, Lite, Pro, and Premier—alongside Canvas for image generation and Reel for video creation.
In benchmarking tests, Nova Pro goes toe-to-toe with heavyweights like GPT-4o, Mistral Large 2, and Llama 3, often coming out on top.
These models boast some impressive stats, supporting over 200 languages and handling context windows up to 300,000 tokens, with ambitions to hit 2 million tokens by 2025.
The Reel model already churns out six-second videos from text or image prompts, with plans to stretch its creative muscle to two-minute clips in the near future.
Looking ahead, speech-to-speech and "any-to-any" modality models will join the Nova family by 2025.
Why is this a big deal?
Though Amazon may have arrived late to the AI party, it’s crashing in style. With a colossal user base, deep pockets, and a lineup of models that are no slouches in the performance department, Amazon might just sneak up the AI leaderboard and give the incumbents a serious run for their money. Dark horse, anyone?
Also in AI
Design
Running with Hypes and Trends is Not Always a Good Trade-Off
The visual world’s been adapting to the smartphone’s lens, where instant, shareable moments reign. Hotels, restaurants, and events now craft Insta-ready settings to lure influencers, with ‘going viral’ driving consumer culture.
Prime’s Flash in the Pan
Roope contrasts Prime’s rapid hype-driven growth with Tenzing’s steady, substance-based approach. While Prime exploded with $250 million in its first year, it’s now stacked in discount stores, a cautionary tale of a brand without lasting value.
Tenzing’s Winning Formula
Unlike Prime, Tenzing values simplicity, authenticity, and quality. The energy drink brand focuses on building real relationships, steering clear of influencer-led hype, and sticking to values that resonate with customers.
No gimmicks, just a solid product and brand purpose.
The Takeaway
While hype may burn bright, it’s short-lived.
Tenzing’s approach is the long game—substance, trust, and purpose matter more for sustained growth. And while it’s not flashy, its design stays true to its message, something that could one day go viral on its own merit.
In short, stay grounded—hype might be fun, but substance sticks.
Also in Design
Science & Tech
Brain Mapping Using Viruses: Creating Neuron "Barcodes"
In October, scientists achieved something remarkable—a complete map of a fruit fly’s brain.
With around 140,000 neurons and 500 feet of wiring, the map revealed complex cell types and connections. This marks a major leap in understanding brain function.
An Immense Challenge
Despite the success, the road ahead is daunting. The map of the fruit fly’s brain took hundreds of scientists a decade to compile.
They sliced the brain into thin pieces, photographed them in high detail, then stitched millions of images together using AI and human analysis.
History of Brain Mapping
The previous benchmark was set in the 1980s with a worm's connectome of just 302 neurons. Moving from that to the fly’s brain with 140,000 neurons required major scientific breakthroughs.
In short, while the fruit fly’s brain map is a huge achievement, it also highlights the immense difficulty of advancing brain mapping to more complex organisms.
Also in Science & Tech
Meta Eyes Nuclear Power for Data Centres: Aiming for sustainable AI-driven infrastructure
Student Rocket Group Breaks Records: USC's Aftershock II rocket hit 470,000 feet
Amazon Aurora DSQL: A Serverless, Multi-regional, and Postgres-compatible Cloud Database
Samsung Smart Glasses: Preview in January alongside the Galaxy S25; release expected Q3 2025
Founding
Product-Market Fit isn’t a one-time task: It Needs To Be Maintained Or Your Numbers Die
Product-Market Fit (PMF) isn’t a one-and-done affair; it’s a never-ending workout. Neglect it, and your business risks losing its edge.
The ZIRP era (Zero Interest Rate Policy) let many companies coast with weak PMF, but today’s tougher environment demands relentless effort to maintain it.
Why Companies Fold
Startups don’t run out of money—they run out of PMF. Mismanagement of cash or bungled execution only hastens the inevitable if PMF isn’t strong.
Weak PMF equals inevitable collapse, no matter how robust your financial planning or operations may seem.
What PMF Really Means
PMF is about creating a product so compelling that customers rave about it without prompting. In Sam Altman’s words: “Do any users love our product so much they spontaneously tell other people to use it?”
It’s a living concept that needs constant nurturing—especially as market needs evolve.
The Dangers of False Confidence
Unicorn companies with bloated valuations often mistake hefty funding for proof of PMF. It’s not.
Huge investments can create unsustainable expectations, forcing premature scaling and neglect of PMF. It’s not "death by funding," but "death by delusion."
Winning Strategies for PMF in 2025
The game has changed. Specialised, vertical SaaS and AI-first solutions outperform broader offerings.
Compound startups—offering integrated solutions—are thriving as customers demand more value from fewer providers.
Remember: whether you’re scaling a unicorn or launching a startup, PMF isn’t a milestone. It’s your business’s heartbeat—ignore it at your peril.
Also in Founding
Product
Pre-mortem Meetings Used by Meta and Instagram: A Pre-mortem of Product Before Launch
Why wait for the shipwreck when you can chart a safer course upfront? Here’s how Meta, Instagram, and other top tech teams tackle challenges before they even appear.
What’s a Pre-Mortem, Anyway?
Instead of dissecting what went wrong post-disaster (classic post-mortem), a pre-mortem flips the script. It’s like saying, “The launch tanked. What tripped us up?” before you actually hit the iceberg.
In this no-holds-barred session, the team lays out their worries, plots out risks, and nips potential disasters in the bud. It’s all about boosting the odds of a smooth launch and avoiding those pesky project calamities.
Aligning the Crew: Pre-Mortems as Your Team’s Compass
Sure, you’ve got project kick-offs and product reviews—but plans change, markets shift, and sometimes, folks aren’t on the same page. Enter the pre-mortem: a sneaky good way to pause, align, and make sure the ship’s still pointed towards success.
Research says imagining failure boosts your ability to predict it by 30%. Handy, right? Plus, it keeps everyone from engineering to legal in sync, so no one’s left scrambling when the big day rolls around.
Running a Pre-Mortem: From Tigers to Elephants
Here’s the playbook:
Assemble Your Crew: Beyond the usual suspects (PMs, engineers), bring in marketing, legal, and even customer support. Diversity of thought is key.
Set the Scene: Imagine launch has failed. Now, brainstorm every reason it could’ve gone belly-up. Use these categories:
Tigers: Real threats ready to pounce.
Paper Tigers: Worries that look scary but aren’t.
Elephants: Concerns no one’s talking about but should.
Get Everyone Talking: Use Post-it notes, group themes on a whiteboard, and keep it safe—no idea is too silly.
Sort the Issues:
Launch-Blockers: Fix now.
Follow-Ups: Tackle in the next iteration.
Trackables: Keep an eye on these.
Assign Ownership: Every Tiger gets a handler, even if it’s just someone to find the real owner.
Want fewer flops and more wins? Pre-mortems are your golden ticket. Get your team on board, brainstorm those Tigers, and you’ll be sailing smooth before you know it.
Today’s AI Image
Amazon’s AI Family
Source - DALL-E
Quote of the Day
A Vital Life Principle from Orwell
"In a time of deceit telling the truth is a revolutionary act."
George Orwell
What we’re working on
Velvet Onion & Friends We’re in the process of rebranding Velvet Onion & Friends. Why? It’s an important stage in our evolution, and deepens the link between agency, product & education. | We’re at the final stages of planning for our pilot program. Working name is “99 Problems But A Pitch Ain’t One;” cute for internal projects, not sure it’s the name. Coming soon! |