December AI Surge: Robots, Gemini, Claude
Key Points
- A new humanoid robot built by robotics firm Abtronic in partnership with Google DeepMind aims to give AI real‑world sensory data, which could help overcome the “pre‑training wall” and enable intelligence to scale beyond internet‑derived data.
- Google released Gemini 2.0 “experimental thinking,” a model that outranked OpenAI’s GPT‑4 on leaderboards, delivering detailed critiques, rewrites, and human‑level intent explanations that make it useful for final‑draft content generation.
- Claude (Anthropic) announced a long‑awaited Excel‑file understanding update, allowing the model to ingest and manipulate structured spreadsheets up to roughly 30 MB, addressing a major limitation for LLMs working with tabular data.
- Rumors suggest OpenAI will unveil a new model dubbed “03” later today, following a pattern of frequent releases (e.g., the “12th day of Christmas” analogy) that may further intensify competition among the leading AI providers.
- These four developments—robotics integration, Gemini 2.0’s performance, Claude’s Excel capability, and the upcoming OpenAI release—highlight a rapid escalation in both AI hardware and software capabilities as the industry races toward more versatile, real‑world‑aware intelligence.
Full Transcript
# December AI Surge: Robots, Gemini, Claude **Source:** [https://www.youtube.com/watch?v=izfgyzADwKo](https://www.youtube.com/watch?v=izfgyzADwKo) **Duration:** 00:05:54 ## Summary - A new humanoid robot built by robotics firm Abtronic in partnership with Google DeepMind aims to give AI real‑world sensory data, which could help overcome the “pre‑training wall” and enable intelligence to scale beyond internet‑derived data. - Google released Gemini 2.0 “experimental thinking,” a model that outranked OpenAI’s GPT‑4 on leaderboards, delivering detailed critiques, rewrites, and human‑level intent explanations that make it useful for final‑draft content generation. - Claude (Anthropic) announced a long‑awaited Excel‑file understanding update, allowing the model to ingest and manipulate structured spreadsheets up to roughly 30 MB, addressing a major limitation for LLMs working with tabular data. - Rumors suggest OpenAI will unveil a new model dubbed “03” later today, following a pattern of frequent releases (e.g., the “12th day of Christmas” analogy) that may further intensify competition among the leading AI providers. - These four developments—robotics integration, Gemini 2.0’s performance, Claude’s Excel capability, and the upcoming OpenAI release—highlight a rapid escalation in both AI hardware and software capabilities as the industry races toward more versatile, real‑world‑aware intelligence. ## Sections - [00:00:00](https://www.youtube.com/watch?v=izfgyzADwKo&t=0s) **Google’s New Humanoid Robot Initiative** - The speaker outlines Google DeepMind’s partnership with robotics startup Abtronic to launch a humanoid robot, stressing how real‑world sensory data from such bots could break the pre‑training ceiling and accelerate the scaling of AI intelligence. ## Full Transcript
today December 20th is already a massive
day in AI before open AI drops whatever
they're dropping for the 12th day of
Christmas and the rumor is it's
something called 03 we will see but
already we have four big developments
for you number one abtronic which is a
robotics company has collaborated with
Google Deep Mind to release a humanoid
robot powered by Google AI this is big
from a long-term perspective because one
of the significant bets on the ability
of AI to continue to scale intelligence
is that we find another big data pool so
when Ilia suer talked about the internet
being our biggest data pool and it's
being something that is not renewable
the only way around that is if you give
AI access effectively to situational
awareness of the real world Elon has
called that out with his cars and his
robots in this case Google's entering
that Arena as well the idea is that when
humans learn human babies take in far
more data in their eyes and their ears
than we give to even our largest large
language models in their first three
four years of life if you can give a
robot that kind of data input from
interacting with the real world as you
would if it was a you know actual
walking around robot well maybe that's a
way to get through the pre-training wall
and start to continue to scale
intelligence so that's the Strategic
reason why Google getting into the
robotic space is so
interesting but Google's not done yet so
Google also in the last 24 hours
released a model that took the top spot
in the leaderboards from open AI
01 it's called Gemini 2.0 experimental
thinking it's available in the AI studio
now from Google I have played with it it
is amazing
I gave it something that was a document
that I thought was okay that uh Claude
Sonet 3.5 had written and I said can you
make this better it was a very short
prompt I did not do my best job at
structuring a prompt it came back with
the most detailed critique of how to
make the doc better rewrote the entire
Doc and described it in a way I could
understand and then gave me the human
intent behind it like the reason why it
did what it did in a way that a human
could understand and it made a ton to
sense
I was shocked I ended up using Claude
for
formatting and that's not really what
you're supposed to use a large language
model for but this model was so good
Gemini thinking was so good that I just
didn't need to touch it like we've gone
from like it can be a draft to this
might be a final draft and that's a big
step forward and again open a may drop
something even better later today we
will see so that's number two Gemini 2.0
thinking check it out
number three Claude released a long
awaited update to excel understanding so
Excel file understanding has been a huge
issue for large language models Claude
has been at the Forefront of using tool
sets to understand these structured data
sets and these tables and Claude
released a update that
essentially the anthropic team is going
to let you handle a Excel file up to 30
Megs in size larger than the normal
context window and they're not really
clear quite how they do it but at the
end of the day Claud is going to be able
to look across that entire spreadsheet
and extract meaningful insights even if
it exceeds the traditional definition of
the context window and that may be as
simple as they're adding a special
context window that they can trigger
when a very large Excel file goes in but
it's still significant because
structured data sets increase
combinatorially in complexity the bigger
they get and so a 30 megga Excel file
like I've worked with those Excel files
they're really really hard to understand
for a human and so getting the ability
to like pull that into an AI is a big
step
forward
finally uh I did not know this until
today but apparently AI is getting to
the point where it can pass the mirror
test so one of the classic tests for
intelligence in the animal kingdom is
can an animal recognize that an image in
the mirror is itself my Corgi cannot do
this my Corgi is dumb as a sack of
hammers but there are animals that can
do this gorillas can do this there are
other animals that can do this as well
it's a well-known test in biology and so
of course people are wondering can AI do
this and the answer is AI is getting
better and better and better at this
Claude has passed the self-awareness
test now the mirror test if you take a
screenshot of
Claude and give it to Claude that's how
you do the mirror test and Claude can
pass that now it doesn't mean that
Claude passes it as often as humans do
uh they apparently have a benchmark for
self-awareness I didn't even know this
and AI uh the four class models score at
about 50% on self-awareness and humans
score above 90% I didn't know we didn't
score 100% but I guess we don't uh maybe
nothing is 100% in this world but
anyway the point is that the four class
models are significantly better than the
three class models at self-awareness and
we should expect them to continue to get
better and yes that imposes really deep
philosophical questions and we're going
to be asking a lot more philosophical
questions around AI in 2025 so that's
your update I will drop something else
later in the day uh as open AI has their
final day release party but I thought
this news was too important not to share