STARTUPS PULSE - Reinvent The Wheel

Meta Invests Billions in Scale AI CEO to Boost AI Capabilities Coco Robotics Raises $80 Million, Partners with DoorDash and Uber Eats #startups, #AI, #Meta, #CocoRobotics, #investments, #DoorDash, #UberEats

Episode Transcript

(00:00):
Welcome to Innovation Pulse, your quick no-nonsense update covering the latest in startups and

(00:09):
entrepreneurship news. Scale AI is attracting Meta's interest for a potential $15 billion
investment, while Coco Robotics secures $80 million in funding, advancing urban delivery
tech. After this, we delve deeper into the intriguing balance between reinvention and
reuse in tech innovation. Scale AI is a data labeling and annotation startup founded in

(00:36):
2016 that has become a key player in the generative AI era. Its core service involves helping tech
giants like OpenAI, Google, and Microsoft prepare data to train advanced AI models. The startup caught
Meta's attention, prompting a potential investment of up to $15 billion for a 49% stake, as Meta

(01:01):
aims to enhance its AI capabilities. Scale AI is unique due to its expertise in data
preparation for AI models, a critical step for any AI development process. This expertise has led to
collaborations with major tech companies and government entities like the US Department of

(01:22):
Defense. Additionally, Scale AI's collaboration with Meta on Defense Lama, an AI model for national
security, underscores its strategic importance. The startup's value lies in its ability to
simplify complex data processes for AI training, making it indispensable for companies aiming to

(01:43):
develop cutting-edge AI solutions. Meta's investment signifies a strategic move to bolster
its AI ambitions by leveraging scale AI's expertise, especially given its recent challenges in the AI domain.
Coco Robotics, officially known as Cyan Robotics Inc. has secured $80 million in new funding,

(02:08):
with significant backing from OpenAI CEO Sam Altman. This brings their total fundraising to
over $110 million. Established in 2020, Coco Robotics operates a fleet of approximately 1,300
electric delivery robots across cities like Miami, Chicago, Los Angeles, and Helsinki.

(02:31):
The company's robots are cooler-sized and designed to navigate urban environments,
enhancing food delivery services. Coco Robotics has formed strategic partnerships with major
players like DoorDash and UberEats and is also collaborating directly with merchants.
A notable alliance with OpenAI allows Coco to integrate advanced navigation technology,

(02:56):
enabling its robots to efficiently maneuver around obstacles and make real-time decisions.
The data shared with OpenAI helps refine urban delivery logistics, providing a unique edge in
robot-assisted deliveries. This combination of cutting-edge technology and strategic partnerships
positions Coco Robotics as a valuable innovator in urban delivery solutions, offering an efficient,

(03:22):
scalable, and tech-driven alternative to traditional delivery methods.
And now, pivot our discussion towards the main entrepreneurship topic.
All right, everybody. Welcome to another deep dive on innovation pulse.

(03:42):
I'm Donna, and as always, I've got my co-host, Yaakov Lasker here with me.
Today, we're tackling something that might make some of you engineers and builders out there
squirm a little bit in your seats. Oh, this should be good. Lay it on me, Donna.
We're talking about one of the most sacred pieces of advice in tech and engineering.
Don't reinvent the wheel. And I've got to tell you, Yaakov,

(04:04):
I'm starting to think this advice might actually be doing more harm than good.
Wait, hold up. Are you seriously about to argue that we should be reinventing wheels left and
right? Because I've seen what happens when junior developers decide they're going to build their own
JavaScript framework instead of using React. Exactly. See, you just proved my point. You

(04:26):
immediately went to that defensive place that so many of us go to. But here's the thing. Think
about who typically gives this advice. It's usually people who either try to build something
themselves and know how brutally hard it is, or people who never tried at all and are just
repeating what they've heard. Okay, I'm listening. But you've got to admit there's wisdom in not

(04:47):
wasting time rebuilding something that already works perfectly well. But does it work perfectly well?
Think about it this way. We've had wheels for over 6,000 years since around 4500 BCE.
If everyone had followed the don't reinvent the wheel advice, we'd still be rolling around on
solid wooden discs. We wouldn't have pneumatic tires. We wouldn't have ball bearings. We wouldn't

(05:12):
have any of the incredible wheel innovations that make modern transportation possible.
That's actually a fascinating point. The wheel itself has been reinvented countless times across
different civilizations and cultures. And each reinvention brought something new to the table.
Exactly. And here's where it gets really interesting for us in the tech world.

(05:34):
Richard Feynman, the Nobel Prize winning physicist, had this incredible quote,
What I cannot create, I do not understand. Think about what that means for a second.
So you're saying that to truly understand something on a fundamental level, you need to be able to
build it yourself. Even if it's just a toy version. Bingo. And this is where I think we've been doing

(05:56):
ourselves a huge disservice. In computer science, we treat so many things like they're beyond the
reach of mere mortals, protocols, cryptography, web servers. But here's the secret. Humans,
just like us, created all of these things. They're not perfect. And we can absolutely make different
tradeoffs in our own designs. You know what? I'm starting to see where you're going with this.

(06:19):
I remember when I first tried to implement my own simple HTTP server, just to understand
how they work. It was terrible. Absolutely terrible. But I learned more about networking in those two
weeks than I had in years of just using existing servers. Yes. And that's the beautiful thing about
going down these rabbit holes. Even something as seemingly simple as a string or a file path,

(06:42):
turns out to have infinite complexity when you really dig into it. You start to appreciate
just how many decisions and tradeoffs went into the tools we use every day. Right. And that leads
me to something I've noticed. When you're building something from scratch, you're forced to make
decisions about correctness versus simplicity, functionality versus performance, scalability

(07:05):
versus resource usage. You can't optimize for everything. And that's the humbling experience
right there. Building something that even one other person finds useful is incredibly difficult.
But it also shows you that existing solutions aren't perfect either. They were designed with
specific constraints and use cases in mind, and they might not be the best fit for your particular

(07:27):
problem. So what are some good reasons to reinvent the wheel? I mean, beyond just learning how wheels
work. Oh, there are so many. You might want to build a better wheel for some specific definition
of better. Maybe you need to teach others about wheels or learn about the inventors who came before
you. Maybe you want to be able to fix wheels when they break, or you're helping someone who needs

(07:50):
a very specialized wheel, like a wheelchair wheel versus a car wheel versus a skateboard wheel.
They're all wheels, but they solve completely different problems. Exactly. And here's something
really cool. Maybe you fail at building a better wheel, but along the way, you discover a better
way to test wheels. Or maybe your wheel isn't even meant for transportation at all. You might end

(08:13):
up with a potter's wheel or a steering wheel or a flywheel. Innovation often comes from thinking
outside the box like this. This reminds me of how so many great discoveries happened by accident.
People set out to solve one problem and ended up solving a completely different one. Right. And in
software especially, running little experiments is cheap and quick. You can solve your own problem,

(08:37):
start small, keep it simple, and iterate. The key is not to give up before you have a working version
of what you're trying to explore. But we should probably add a caveat here, right? We're not saying
to ignore the work of others or reinvent things out of pure stubbornness or ignorance. Absolutely.
Study the work that came before you. Understand why decisions were made the way they were.

(09:00):
Reuse where it makes sense. But don't let the fear of reinventing stop you from exploring and
learning. So what's the balance then? When should you reinvent versus when should you reuse?
I love this framework. Reinvent for insight. Reuse for impact. If your goal is to learn and
understand something deeply, go ahead and build your own version. Get your hands dirty. Make mistakes.

(09:24):
Learn from them. But if your goal is to ship something that actually helps people, then by
all means, stand on the shoulders of giants and use the best tools available. That's brilliant.
So reinventing isn't about replacing existing solutions. It's about understanding them well
enough to know when and how to use them effectively. Exactly. And think about it this way. If you never

(09:46):
test your knowledge by trying to implement something yourself, how would you ever learn
enough about your field to advance it? The people who push industries forward are often the ones who
understand the fundamentals deeply enough to see new possibilities. You know what this reminds me of?
The best senior engineers I know aren't the ones who memorize every API and framework.

(10:09):
They're the ones who understand the underlying principles well enough
to adapt to new technologies and make smart architectural decisions.
Yes. Because when you understand how something works at a fundamental level,
you can debug it when it breaks, modify it when your needs change, and recognize when it's the
wrong tool for the job. You become a more thoughtful and effective engineer. So for our listeners who

(10:32):
are maybe feeling inspired to go reinvent some wheels, what would you recommend as a good starting
point? Start with something you use every day, but don't fully understand. Maybe it's how your text
editor saves files or how your web browser renders HTML or how your phone connects to Wi-Fi. Pick
something that genuinely interests you, not something you think you should learn. And remember,

(10:56):
the goal isn't to build the next great commercial product. It's to build understanding.
Your implementation can be messy, incomplete, and full of bugs. That's totally fine. You can throw
it away when you're done learning from it. The beautiful thing is that in our field, experimentation
is relatively cheap. You're not dealing with physical materials or manufacturing costs.

(11:20):
You can afford to try things, fail, learn, and try again. And who knows?
Maybe along the way, you'll discover something that nobody else has thought of.
Or maybe you'll just become a better engineer who makes more informed decisions about when to use
existing tools and when to build something new. The key insight here is that curiosity and exploration

(11:41):
shouldn't be discouraged. We need more people who are willing to dig deep, ask questions,
and challenge assumptions. That's how industries evolve and improve. So maybe it's time to retire
that old advice about not reinventing the wheel. Or at least add some important context to it.
I couldn't agree more. The next time someone tells you not to reinvent the wheel, maybe ask them,

(12:04):
but how else will I learn how wheels work? Sometimes the best way to appreciate existing
solutions is to understand the problems they solve. And remember, every breakthrough innovation in
history started with someone deciding to reinvent something that already existed. They just made
it better, different, or more suitable for a new context. So here's our challenge for everyone

(12:27):
listening. Pick something you use every day, but don't fully understand. Spend a weekend trying to
build a simple version of it. You might be surprised by what you learn, not just about the technology,
but about yourself as a problem solver. Whether you're a seasoned engineer or just starting out,
there's always something new to discover when you roll up your sleeves and start building.

(12:50):
The wheel has been around for 6,000 years and we're still finding new ways to improve it.
Thanks for joining us on this deep dive into the art and science of reinventing wheels.
Remember, reinvent for insight, reuse for impact. Until next time, keep innovating.
And keep those wheels turning, everyone. We'll see you next episode on Innovation Pulse.

(13:12):
As we wrap up today's podcast, we've delved into scale AI's pivotal role in data preparation for
AI giants and cocoa robotics advancements in urban delivery, alongside a discussion on balancing
innovation with practicality in tech. Don't forget to like, subscribe, and share this episode with

(13:37):
your friends and colleagues, so they can also stay updated on the latest news and gain powerful
insights. Stay tuned for more updates.