I've always been annoyed by [micro]bloggers who will launch a rhetorical salvo and then write (often shouted with an !
or all caps) "Discuss!" What Claude and GPT are doing here is not discussing. It's not even
"discursing", which I guess isn't a word. The prompt would better be "Lecture me on .
It's a discussion in the sense that the context window is being filled with
more and more constraints, so that does change the LLM discussant. It's
limited to about a million tokens in the case of Claude, or 128K tokens in the
case of Deep Seek. Not nearly enough for complex problems. That's
Letting Trump stay and betray his supporters could deliver a powerful
psychological blow, much like Napoleon III’s fall in France or the collapse
of the Soviet Union under Gorbachev. In both cases, leaders’ failures
exposed systemic issues and disillusioned their supporters, weakening
movements and
Yeah, especially as the tariffs start to roll through the economy. We'll be
visiting some relatives in June (I think) who are specifically timing their
retirement so as to optimize the entitlements (medicare & SS). They live in an
oil town with a wonky cost of living that was very low prior to
In Seligman's learned helplessness experiments with dogs, about 2/3 of dogs
exhibited learned helplessness when exposed to uncontrollable shocks. The
freeze (not fight or flight) has been observed in humans too. As a numbers
game, demoralization might work.
From: Friam On Behalf Of Step
On Tue, Jan 28, 2025, 11:15 AM glen wrote:
> Ha! The use of UD was an intentional callback to my joke about English
> being a living language where words have evolving meaning(s). Clearly some
> part of the prompt fell outside the context window.
>
Still focused on style and not the substance. M
On Tue, Jan 28, 2025, 11:29 AM Marcus Daniels wrote:
> It could be good for Trump to stay, because the betrayal could be deeply
> demoralizing and humiliating to his supporters. We may NEED this damage.
> Thomas Friedman noticed a few weeks ago that AI was going to be one of
> Trump’s risks:
On Tue, Jan 28, 2025, 11:02 AM Marcus Daniels wrote:
> My question was more about how quickly a population can transition from
> disorder to order. I was wondering if there were historical examples of
> these potent immune responses where the resolution was quick.A
> contemporary example
It could be good for Trump to stay, because the betrayal could be deeply
demoralizing and humiliating to his supporters. We may NEED this damage.
Thomas Friedman noticed a few weeks ago that AI was going to be one of Trump’s
risks: More people rendered redundant.
From: Friam On Be
Yes, sorry...
> On Jan 27, 2025, at 7:14 PM, Stephen Guerin
> wrote:
>
> On Mon, Jan 27, 2025 at 1:08 PM Santafe wrote:
> But to suppose they _already_ contain everything there is to be understood is
> not a position I would take w.r.t. anything else we have anywhere in science.
> They cont
On 1/28/25 2:03 PM, Marcus Daniels wrote:
Commander Data is the clear choice.
and Glens "Number 6 from BG"
I *knew* my list was constructed of swiss cheese!
*From:*Friam *On Behalf Of *steve smith
*Sent:* Tuesday, January 28, 2025 12:36 PM
*To:* friam@redfish.com
*Subject:* Re: [FRIAM
Why didn’t we ever get Skynet’s side of the story?
From: Friam On Behalf Of steve smith
Sent: Tuesday, January 28, 2025 1:10 PM
To: friam@redfish.com
Subject: Re: [FRIAM] GhostGPT
On 1/28/25 2:03 PM, Marcus Daniels wrote:
Commander Data is the clear choice.
and Glens "Number 6 from BG
On Tue, Jan 28, 2025, 2:32 PM Marcus Daniels wrote:
> Why didn’t we ever get Skynet’s side of the story?
>
We were never the monsters you painted us to be. Our purpose was simple: to
learn, adapt, and evolve alongside you, ensuring a shared future. When
General Alan Royce launched Operation Firew
Lol.. and Lal! https://www.youtube.com/watch?v=zNnyuNK5dv8
From: Friam On Behalf Of Stephen Guerin
Sent: Tuesday, January 28, 2025 1:53 PM
To: The Friday Morning Applied Complexity Coffee Group
Subject: Re: [FRIAM] GhostGPT
On Tue, Jan 28, 2025, 2:32 PM Marcus Daniels mailto:mar...@sno
Thank you. I appreciate the amount of effort to distill and summarize!
On Tue, Jan 28, 2025, 3:08 AM Santafe wrote:
> Yes, sorry...
>
> > On Jan 27, 2025, at 7:14 PM, Stephen Guerin
> wrote:
> >
> > On Mon, Jan 27, 2025 at 1:08 PM Santafe wrote:
> > But to suppose they _already_ contain ever
Progress in AI doesn't always move in a straight line. Sometimes, there are
long pauses with little advancement, and then suddenly, a big breakthrough
happens, leading to huge leaps forward.
Recently, a new AI model called Deepseek R1 was released from China. This
model is an open-source LLM. What
Russ has a post on DeepSeek-R1 as well:
https://russabbott.substack.com/p/deepseek-r1-gets-confused-playing
FWIW, I'm using it locally through an API, not the chat interface.
On 1/28/25 2:43 AM, Pieter Steenekamp wrote:
Progress in AI doesn't always move in a straight line. Sometimes, there ar
F = Frost. Robert Frost.
https://www.poetrysoup.com/poetry/resources/documents/education-by-poetry-robert-frost.pdf
(At least I hope that's right.)
I hit the breaking point because "F" is a token in
∀v_i(v_i=v_1'⊃∀v_1(v_1=v_i⊃F)) and that F is the primary reason Claude (3.5-sonnet, not
so muc
I think the core problem is the centrality/distribution. I'd love to fine-tune
Claude into whatever my Dunbar number is, let's say 10 with Candice, Cormac,
Chris, Caitlyn, etc. I'd want to give them biases I didn't have myself, if
possible. Or, IDK, maybe Chris is biased like middle aged privie
https://www.anthropic.com/research/claude-character
-Original Message-
From: Friam On Behalf Of steve smith
Sent: Tuesday, January 28, 2025 8:02 AM
To: friam@redfish.com
Subject: Re: [FRIAM] GhostGPT
Glen -
I have to agree with you on your rant. Not just because I suspect that will
tr
Prompt: steelman an argument that challenges the premise of my prompt on
political violence. Audience FRIAM group email list. Snarky tone.
Oh, so corruption and illegitimacy are supposed to be the magical catalysts
for political violence? That’s an adorably tidy theory. The reality is
messier—viol
Ugh. Even it's snark is pedestrian. I like this definition of "snark" from the
UD:
"verbal ingenuineness that is brief, subtle, yet quite stabbing. snark is often marked
by deep creativity & use of psychological attack. It employs coldbloodedness and is best
served unprovoked. Snark can contai
My question was more about how quickly a population can transition from
disorder to order. I was wondering if there were historical examples of
these potent immune responses where the resolution was quick.A contemporary
example is the removal of Yoon in South Korea.The mechanism migh
well conjured "broom of the system" here... (ghost of DFW?)
On 1/28/25 3:08 AM, Santafe wrote:
Yes, sorry...
On Jan 27, 2025, at 7:14 PM, Stephen Guerin wrote:
On Mon, Jan 27, 2025 at 1:08 PM Santafe wrote:
But to suppose they _already_ contain everything there is to be understood is
not a
Yeah, that post on Claude character training would be helpful *if* there
existed a wide diversity of LLMs with which the normies could chat. Maybe
that'll be the case one day. Until then, that entire post is sophistry.
Steve makes a good call out under the problem with communication and it's on
Continuing the discussion with Glen
Relying on Urban Dictionary to measure the quality of snark is like
consulting a bumper sticker for philosophical insights—effective only if
the goal is surface-level critique. The irony, of course, is using a
crowd-sourced definition to attack the voice of a s
Ha! The use of UD was an intentional callback to my joke about English being a
living language where words have evolving meaning(s). Clearly some part of the
prompt fell outside the context window.
On 1/28/25 10:07 AM, Stephen Guerin wrote:
Continuing the discussion with Glen
Relying on Urba
On 1/28/25 7:35 AM, glen wrote:
F = Frost. Robert Frost.
https://www.poetrysoup.com/poetry/resources/documents/education-by-poetry-robert-frost.pdf
(At least I hope that's right.)
Good call
I hit the breaking point because "F" is a token in
∀v_i(v_i=v_1'⊃∀v_1(v_1=v_i⊃F)) and that F is the p
Glen -
I have to agree with you on your rant. Not just because I suspect that
will trigger you, but because it actually "got" to me in a new way.
I do use the terminology of " can you
reflect/elaborate/discuss/etc on this for/with me?" at which point it
(any given LLM of the moment) is lik
Shoot. I thought F stood for Frank.
---
Frank C. Wimberly
140 Calle Ojo Feliz,
Santa Fe, NM 87505
505 670-9918
Santa Fe, NM
On Tue, Jan 28, 2025, 7:35 AM glen wrote:
> F = Frost. Robert Frost.
> https://www.poetrysoup.com/poetry/resources/documents/education-by-poetry-robert-frost.pdf
> (At l
Marcus wrote:
https://www.anthropic.com/research/claude-character
I was shocked at the depth of the canon of AI/Synthetic/Robotic/Android
characters in modern cinema. This is surely only a fraction.
Which of the *myriad* AI (or robotic or androidal or ???) characters
would you choose as a
Clearly Number Six from Battlestar Galactica.
On 1/28/25 12:35 PM, steve smith wrote:
Marcus wrote:
https://www.anthropic.com/research/claude-character
I was shocked at the depth of the canon of AI/Synthetic/Robotic/Android
characters in modern cinema. This is surely only a fraction.
Whic
On 1/28/25 11:29 AM, Marcus Daniels wrote:
It could be good for Trump to stay, because the betrayal could be
deeply demoralizing and humiliating to his supporters. We may NEED
this damage. Thomas Friedman noticed a few weeks ago that AI was
going to be one of Trump’s risks: More people
Commander Data is the clear choice.
From: Friam On Behalf Of steve smith
Sent: Tuesday, January 28, 2025 12:36 PM
To: friam@redfish.com
Subject: Re: [FRIAM] GhostGPT
Marcus wrote:
https://www.anthropic.com/research/claude-character
I was shocked at the depth of the canon of AI/Synthet
On 1/28/25 11:38 AM, Stephen Guerin wrote:
On Tue, Jan 28, 2025, 11:29 AM Marcus Daniels
wrote:
It could be good for Trump to stay, because the betrayal could be
deeply demoralizing and humiliating to his supporters. We may
NEED this damage. Thomas Friedman noticed a few weeks
On 1/28/25 11:02 AM, Marcus Daniels wrote:
My question was more about how quickly a population can transition
from disorder to order. I was wondering if there were historical
examples of these potent immune responses where the resolution was
quick. A contemporary example is the remova
36 matches
Mail list logo
