Photo: lilgrapher / Shutterstock.com (Shutterstock)
Elon
Musk
said
his
startup
xAI
will
open-source
Grok
later
this
week,
revealing
the
inner
workings
of
the
chatbot
to
anyone
who
wants
to
dissect
the
AI
or
build
a
competing
product.
The
announcement
—
which
came
just
before
3
am
Monday
morning
in
Texas,
where
Musk
makes
his
home
—
is
short
on
details,
but
it’s
another
poke
in
his
ongoing
battle
with
OpenAI.
The
billionaire
sued
OpenAI
and
its
CEO
Sam
Altman
at
the
beginning
of
March,
arguing
the
company
betrayed
its
mission
to
operate
as
a
nonprofit
and
make
its
research
available
to
the
public.
Musk
was
one
of
OpenAI’s
earliest
investors,
but
he
parted
ways
with
the
company
years
ago,
and
he’s
become
an
increasingly
vocal
critic
after
the
spectacular
rise
of
ChatGPT.
OpenAI did not immediately respond to Gizmodo’s request for comment.
“Open-source”
is
a
vague
term
that
can
mean
anything
from
releasing
every
line
of
code
to
sharing
a
few
key
details,
and
Musk’s
announcement
didn’t
elaborate
on
the
plan.
This
time
last
year,
he
made
an
elaborate
show
over
his
decision
to
take
X
open-source,
though
the
code
he
published
didn’t
reveal
much
beyond
a
high-level
view
of
a
couple
of
the
company’s
algorithms.
Still,
the
CEO
is
a
longtime
proponent
of
open-source
research.
Tesla
has
a
practice
of
releasing
details
about
some
of
its
patents
that
most
companies
would
keep
secret.
Musk isn’t the first to open-source a chatbot. Both Meta and the recent French upstart Mistral AI have published some of their AI code, though in both cases, those approaches keep key details under the belt.
But
the
move
serves
to
strengthen
Musk’s
argument
against
OpenAI.
To
a
large
extent,
Altman
has
abandoned
his
company’s
commitment
to
working
in
public.
Musk
put
it
bluntly
during
his
late-night
tweet
story.
“OpenAI
is
a
lie,”
he
wrote.
Comments