OpenAI Sora Unveiled, Google Willow Chip
Key Points
- OpenAI finally unveiled Sora, its long‑teased text‑to‑video model, but shut down sign‑ups within an hour because the surge in demand outstripped the company’s available compute capacity.
- A recent leak of Sora footage by artists amplified the hype, and while the service currently only produces very short clips (5 seconds on Plus, 20 seconds on Pro), widespread access remains uncertain due to the heavy processing required.
- OpenAI’s compute resources are already stretched across other intensive projects such as the 01 and 01 Pro models, complicating the timeline for expanding Sora’s availability.
- Google introduced Willow, a new quantum chip that claims to scale qubit numbers while simultaneously reducing error rates—a breakthrough that tackles the fundamental scalability challenge in quantum computing.
- Unlike typical annual quantum announcements from Google, Willow stands out because its error‑reduction approach could make larger, more stable quantum systems viable for real‑world applications.
Full Transcript
# OpenAI Sora Unveiled, Google Willow Chip **Source:** [https://www.youtube.com/watch?v=u__f5qh3p0Y](https://www.youtube.com/watch?v=u__f5qh3p0Y) **Duration:** 00:07:00 ## Summary - OpenAI finally unveiled Sora, its long‑teased text‑to‑video model, but shut down sign‑ups within an hour because the surge in demand outstripped the company’s available compute capacity. - A recent leak of Sora footage by artists amplified the hype, and while the service currently only produces very short clips (5 seconds on Plus, 20 seconds on Pro), widespread access remains uncertain due to the heavy processing required. - OpenAI’s compute resources are already stretched across other intensive projects such as the 01 and 01 Pro models, complicating the timeline for expanding Sora’s availability. - Google introduced Willow, a new quantum chip that claims to scale qubit numbers while simultaneously reducing error rates—a breakthrough that tackles the fundamental scalability challenge in quantum computing. - Unlike typical annual quantum announcements from Google, Willow stands out because its error‑reduction approach could make larger, more stable quantum systems viable for real‑world applications. ## Sections - [00:00:00](https://www.youtube.com/watch?v=u__f5qh3p0Y&t=0s) **OpenAI's Sora Launch Overwhelms Users** - OpenAI finally unveiled its long‑teased text‑to‑video model Sora, but a pre‑release leak sparked massive demand that quickly exhausted available compute, limiting access to short clips on the Plus tier. ## Full Transcript
all right we have two pieces of news
today first the Sora announcement which
has been long awaited I think it's 9
months now I thought it was seven or
eight but I I'm wrong time is a flat
circle it's been at least nine months
since Sora was teased and open AI
finally announced it and no you cannot
have it and we'll get into why uh and
then Willow which is arguably bigger
news that Google dropped by surprise so
Sora first Sora is a text to video
system Sora was teased a long time ago I
guess it was 9 months ago and it looked
incredible then there was a leak over
the weekend by some angry artists who
were in a theater who were filming it
secretly and it looked incredible in the
leak so ironically the artist probably
increased uptake on the new product and
uptake was so high after 10: a.m.
Pacific yesterday when they announced it
that within an hour they were shutting
down access they just did not have the
compute to handle it and that suggests
to me that they don't know when most
people who were trying to sign up
through the day going to get access they
severely underestimated demand and I
think that was a foreseeable problem
like at the end of the
day if you tease this for nine months
and then an artist leaks it the weekend
before which you know you can't control
but you should expect a lot of demand
because the leak looked incredible
everyone wants to type in their move the
idea and see it come to life and even
though we're all only getting like 5sec
videos on Plus plans and 20 second
videos on Pro plans we we have this
ability to take our imagination and put
it on a screen and we we all want to
try and I I sympathize with them they've
built something so magical that everyone
wants it like sympathy isn't the right
word it's incredible they have a
wonderful highquality problem I'm in awe
maybe that's a better word like well
done but I don't know when we're going
to access it because this is heavy heavy
compute to get a video out of a piece of
text they have a bunch of other heavy
compute applications they have more days
of open AI they are doing new kinds of
compute with 01 and 01
Pro so I'm not sure what's going to
happen or when Sor is going to be widely
available I know they're doing their
best uh and we'll just have to see the
team is building magic effectively and
it turns out most people in the world
want magic and we will have to see when
the computers are available so that we
can all play with our text video
ideas moving on to Willow arguably the
bigger news of the day Willow is a
Quantum chip from Google Now I want you
to have the context that Google does
Quantum announcements every fall like
clockwork they just do and the reason
why this one is
interesting
is this chip has
demonstrated the ability to scale cubit
while reducing
errors for 30 some years Quantum
Computing has meant scaling Quantum
cubits and getting more and more errors
so the system becomes more unstable it's
inherently unscalable Google thinks they
have solved that scalability problem and
The Willow chip is basically a
demonstration piece to say look we can
do this at a scale of 105 cubits a cubit
by the way is instead of the classical
bit with two states a cubit has three
states it has zero and one and super
position and having that additional bit
state allows it to calculate in parallel
so it can calculate eight states in
parallel with like a single bit for
example now you start to scale that up
you start to get very very combinatorial
very very
fast and so Google rightly called out
that part of why they're focused on
Quantum is they see the potential
long-term for Quantum to outscale AI
and this is where the caveats come in
there was a lot of hand ringing on the
internet that like our passwords aren't
safe and our Bitcoin seed phrases aren't
safe and this and that this is not a
commercially scalable chip this is not
coming to iPhone this is not coming to
laptops this is living in a research lab
it may get to a few science labs that
want to look at Quantum
interactions but Google themselves
admits they're a long way from something
stable enough that they can go after
a potential chip build right now the
cubits themselves world record they're
staying staying stable for five micros
seconds I
believe it it it's a tiny amount of
time and if you are having trouble with
stability and you're still trying to
figure out how to
scale yes like for people who are in the
space of cryptography now is probably
the time to start thinking about
postquantum cryptography
good thing to get ahead on that and
there are people who are based on what I
read on X they're like running to their
computers and working on it but for
ordinary people this is not going to be
used to hack things right away there's
just not enough of them they're not in
that uh environment and even the test
that was run the really famous one where
it was like the the headline on the
announcements they can do a calculation
in five minutes that would take a
supercomputer 10 septian
years amazing but it was a deliberately
designed test calculation it wasn't a
piece of like algorithmic math we use in
computers so it does a test calculation
really fast that's great but that
doesn't really help us with building
software and systems so I don't want to
take away from the achievement because
getting to a spot where you can scale
cuits and reduce errors over time is a
massive achievement it's something that
the quantum field has been working on
for 30 years but that does not mean that
it is commercially available and I saw
some misunderstanding about that one
last piece on the willow thing they
casually threw out there I don't know
another word for it it was an aside in
the willow announcement and I'll link to
that they said we
suspect that the ability of this chip to
solve problems where supercomputers
would exceed the length of the livable
universe and to do so in less than five
minutes is in line with the theory that
Quantum Computing occurs in multiple
universes at once essentially what
Google said is their chip is Computing
in the
Multiverse and uh by the way it's just
you know by the way there's a Multiverse
and we're Computing in it thanks Google
I don't know what to do with that uh and
uh this is a strange timeline to be
living in and we will see how things
progress from here every day seems to
have a lot of news cheers