Before the iPhone, the wireless phone industry could not have been more different than what it is today.
Wireless network providers (carriers) had pretty much all the control.
After the global success of the iPod, Steve Jobs wasn’t used to hearing the word “no.”
But when looking for the iPhone launch partner, that’s all he heard.
He had one last meeting with the only company who hadn't rejected him yet.
Here’s how he negotiated in that meeting 🧶👇
Before the iPhone, the wireless phone industry could not have been more different than what it is today.
Wireless network providers (carriers) had pretty much all the control.
Phone manufacturers had little autonomy to do what they wanted.
Steve, obviously, had other plans.
And, in classic Steve fashion, he wanted to be in control of it all.
So he came up with his list of demands.
- own the full design, manufacturing, and marketing process
- sell the iPhone in its own stores
- control all the software
- receive a revenue split from each iPhone users phone bill
Most importantly, they needed agree to all these terms at a time when the iPhone was just a concept and a couple of sketches.
So what was in it for them?
Being the exclusive partner of a potentially world-changing product, for 5 years.
He first approached Verizon.
Verizon went on to say: "The iPhone product is something we are happy we aren't the first to market with.”
He met with other carrier CEO’s and each meeting ended similarly.
Eventually, Steve met with AT&T.
But AT&T was his last shot at one of the nation’s top carriers, he had to make this work.
“Let’s just give em $1 billion”
Steve wagered he’d cut AT&T a check today, and if the deal didn’t work out in their favor, they could keep all the money.
Rather than accept the semi-serious/crazy proposal, they signed the original contract with Steve and became the exclusive launch partner of the iPhone for 5 years.
More from All
How can we use language supervision to learn better visual representations for robotics?
Introducing Voltron: Language-Driven Representation Learning for Robotics!
Paper: https://t.co/gIsRPtSjKz
Models: https://t.co/NOB3cpATYG
Evaluation: https://t.co/aOzQu95J8z
🧵👇(1 / 12)
Videos of humans performing everyday tasks (Something-Something-v2, Ego4D) offer a rich and diverse resource for learning representations for robotic manipulation.
Yet, an underused part of these datasets are the rich, natural language annotations accompanying each video. (2/12)
The Voltron framework offers a simple way to use language supervision to shape representation learning, building off of prior work in representations for robotics like MVP (https://t.co/Pb0mk9hb4i) and R3M (https://t.co/o2Fkc3fP0e).
The secret is *balance* (3/12)
Starting with a masked autoencoder over frames from these video clips, make a choice:
1) Condition on language and improve our ability to reconstruct the scene.
2) Generate language given the visual representation and improve our ability to describe what's happening. (4/12)
By trading off *conditioning* and *generation* we show that we can learn 1) better representations than prior methods, and 2) explicitly shape the balance of low and high-level features captured.
Why is the ability to shape this balance important? (5/12)
Introducing Voltron: Language-Driven Representation Learning for Robotics!
Paper: https://t.co/gIsRPtSjKz
Models: https://t.co/NOB3cpATYG
Evaluation: https://t.co/aOzQu95J8z
🧵👇(1 / 12)
Videos of humans performing everyday tasks (Something-Something-v2, Ego4D) offer a rich and diverse resource for learning representations for robotic manipulation.
Yet, an underused part of these datasets are the rich, natural language annotations accompanying each video. (2/12)
The Voltron framework offers a simple way to use language supervision to shape representation learning, building off of prior work in representations for robotics like MVP (https://t.co/Pb0mk9hb4i) and R3M (https://t.co/o2Fkc3fP0e).
The secret is *balance* (3/12)
Starting with a masked autoencoder over frames from these video clips, make a choice:
1) Condition on language and improve our ability to reconstruct the scene.
2) Generate language given the visual representation and improve our ability to describe what's happening. (4/12)
By trading off *conditioning* and *generation* we show that we can learn 1) better representations than prior methods, and 2) explicitly shape the balance of low and high-level features captured.
Why is the ability to shape this balance important? (5/12)
