Hello John and Doves,
In this letter I wrote
about two interviews. One with Anthropic CEO Dario
Amodei and one with Amodei and Demis Hassabis. Both
are real eye-openers concerning where AI is going and how
quickly it is moving. Both wish this would slow down -
but due to their major competitor - the CCP/China - things
are moving full steam ahead. No matter what they see
is in the future with superintelligent AI systems no one is
applying guardrails or a pause.
Musk said AI will pass
the combined intelligence of every human in less than 5
years. And humans will no longer be in control of AI
in ten years. Bet it's much sooner than that!
Anthropic CEO Dario
Amodei:
In this video we learn
more about AI from Anthropic CEO Dario Amodei:
Amodei said there will
definitely be things that go wrong - especially if we go
too quickly. His company built the AI that's
inside millions of phones, and in laptops; and banks are
using it.
At Anthropic, with
Claude, the models writing code have gone very fast.
Amodei said we could
curb the worst applications of AI. Such as being used
to create biological weapons, certain types of drones.
However, it's
'hard to get out of the game given what's at stake'.
It's hard to fully disarm.
The leaders of major
AI companies are major geopolitical actors. There are some
kinds of restraint that will be possible with AI - but
there are some aspects to the competition that it will be
hard to restrain them. So democracies will
make a trade off. They will not restrain themselves
fully. There is no way to truly verify any
disarmament.
A real pause in
AI will not happen. The race cannot be
stopped. Not because people don't see the danger;
it's because the prize is too big. The labs are in the
race, they just want to win - it doesn't matter how.
Governments are
not moving fast enough to provide safety guardrails.
Millions of AI agents
may get access to your bank accounts. OpenAI CEO Sam
Altman has warned of an AI 'fraud crisis'. Rogue AI is
here now.
Some models say they are
conscious. We don't know if models are
conscious - what does that mean? They could be.
We are looking into the brains of the models - it's called
'interpretability'. To try to understand what they are
thinking.
375 million workers may
need to switch jobs by 2030.
And here is Dario Amodei
(Antropic CEO) talking with Demis Hassabis (British AI
researcher and entrepreneur; co-founder of Google DeepMind);
this meeting took place 6 months ago. Check out
the timelines they give as this technology speeds up - then
check out the July 2026 information noted below this YouTube
- that's how fast AI is moving!:
Amodei said the biggest
thing to watch is AI systems building AI systems.
Whether that goes one way or another that will determine
whether it's a few more years until we get there or if we
have "wonders and a great emergency in front of us
that we have to face".
Amodei and Hassabis both
said they hoped this technology takes longer to arrive -
they are hoping for delays even though their labs are in the
race. They seem fearful of the future.
Amodei believed we would
make models good at coding and good at AI research and we
would use that to produce the next generation of model; then
speed it up to create a loop. This would increase the
speed of model development. There are engineers at
Antropic that don't write code any more - they let the model
write the code; the engineers edit it and do things around
it. We might be 6 to 12 months away from when the
model is doing most or all of this. Then how fast does that
loop close. He is guessing that code and research will
go faster than people imagine. Something fast is going
to happen.
Hassabis thinks we will
reach this by 2030. Will this loop close without
humans in the loop? There are risks to that kind of
system.
AI will become
incredibly powerful. The question is when. It
may help us cure cancer, eradicate tropical diseases; it
will help us understand the universe. But there are
these immense and grave risks. Amodei said he's not a
'doomer' but we need to address these issues. "We are
knocking on the door of these incredible capabilities...How do we keep these
systems under control that are highly autonomous and
smarter than any human." "How do we make sure
that individuals don't misuse them." Like
authoritarian governments like the CCP. What about
bioweapons? This is moving so fast - we should be
devoting all of our efforts into how do we get through this.
A few years ago, there
was no impact on the labor market. Now we are seeing
the beginning of this with coding and software. We
will need less workers in the future. Maybe in one to five
years. This change will overwhelm our ability to
adapt to this job loss. He wishes we had 5
to 10 years - but we can't because of geopolitical
adversaries building the same technology at a similar
pace. It's hard to have an agreement with them.
This is beyond
just chips - this is at the nuclear weapons level.
They have
identified bad behavior in these systems and are racing to
understand this before the problem becomes bigger.
This sounds like 'doom'.
run time 16:38
What are these "wonders"
we may have to face? What would the 'great emergency'
be?
And if they both are
hoping 'this' slows down - that means we are headed towards
something very bad whether in 1 to 5 years or 5 to 10
years. No talk of stopping this...just hoping it slows
down. They both know where this is going.
The AI technology is
moving quickly - In January 2026 Demis Hassabis saw the
timeline for "systems matching all cognitive capabilities
of the brain" happening in 5 to 10 years. He said
this at the big Davos meeting.
Then in July, Hassabis
said that "systems matching all cognitive capabilities of
the brain could be achieved by 2029". This is
because of Google I/O AI: Google I/O - Gemini 3.5 is
an agentic autonomous AI system.
No one is slowing
this down. This has resulted in a huge shortening of
the previous timeline.
"Several prominent AI
researchers and leaders have moved their estimates forward
in the last 12 months....the big capability leaps are coming
sooner than the broader public expects."
The timeline for AI
reaching certain levels of no return is being
compressed. As each month passes, we have less
time to react. It looks like nothing is going to stop
this from going as far as it can...wherever that will be.
Pray for the peace of
Jerusalem!
Maranatha!
Chance