# Source transcript — screenwalk 2610291154

Source: https://youtube.com/watch?v=BMCRxsAthSI&feature=shared

Provenance: YouTube original-English automatic captions (en-orig), retrieved 2026-10-29. Caption text and start timestamps are preserved in order. Not a human-certified verbatim transcript: proper names, specialist and Indigenous-language terms, and speaker attribution need checking. Both human commentary and read-out passages are included; automatic caption markers do not establish an independently participating speaker. YouTube metadata duration is 2160 seconds; the last spoken caption begins at 35:58. The author-supplied description is preserved separately in intake.md, not inserted as transcript speech.

## Automatic caption transcript

[00:00] Okay, I'm going to have a look at my um
[00:06] that's a bigger ceremonia I thought and
[00:09] there's a
[00:11] few session involved. Um and one is
[00:15] doing something interesting.
[00:18] Uh it has um
[00:23] it will be interesting to actually go
[00:25] over
[00:27] what I how I prompted it and what I gave
[00:29] it. Uh so this is a potential
[00:33] output um
[00:36] of the new homepage.
[00:39] uh this um yes we're going to be capable
[00:43] of speaking when we uh query the memory.
[00:48] It's kind of this uh ancient being that
[00:52] doesn't have much color for now. There's
[00:55] an opening here for uh designing inner
[00:58] user interface with uh glyph colors uh
[01:02] and everything but we are focusing on
[01:04] the uh functional aspect here which is
[01:07] uh being capable of speaking
[01:10] uh for your turn
[01:13] uh as well as quering the memory. So we
[01:15] are uh inside a talking circle ceremony
[01:19] there. Uh and here we were and in the
[01:23] homepage where um you can actually
[01:27] there's a bunch of new feature that are
[01:28] getting explored um and put into place.
[01:33] Uh it shows you where you sit presently.
[01:37] It enables you to hask uh the circle's
[01:40] memory. Uh and it also show you it's a
[01:45] the beginning of something that would
[01:47] look like the walls of a a community.
[01:52] Uh and here we are when you actually
[01:54] open a talking circle. So you will be
[01:58] able to speak inside of a talking circle
[02:01] if that makes sense. Uh so you're going
[02:03] to be able to record yourself uh instead
[02:06] of just typing or you have the option of
[02:09] typing uh for your turn for asking the
[02:13] memory which I pretty cool feature. You
[02:16] can actually ask this ceremony this
[02:19] circle which contains all the ceremonies
[02:21] that the circle had and you can oh this
[02:26] is new. you can uh query the prop
[02:30] decomposition that had occurred inside
[02:32] of the actual memory which is very
[02:35] interesting. We spoke about that last
[02:38] week. So, I'm glad to see that there's a
[02:39] the beginning of the implementation for
[02:41] episode 44 cuz uh we really had some u
[02:46] some of the media reviewed yesterday
[02:50] uh which was a very interesting uh
[02:52] aspect and I think I'm going to not
[02:54] necessarily drift but um go in there a
[02:58] little bit. Um and these are uh that's
[03:03] the first uh part of it.
[03:07] I'll present you the foundation of our
[03:09] decomposition service. It helped us
[03:12] start from our intentions toward a
[03:14] structured action plan rooted in the
[03:16] study.
[03:18] >> Okay, we have everything uh line up
[03:21] here. I didn't remember it was one of
[03:23] the exportation that I had with
[03:24] everything. So, we'll listen to it
[03:26] again.
[03:27] >> I'll present you the foundation of our
[03:28] decomposition service. It helped us
[03:31] start from our intentions toward a
[03:33] structured action plan. Rooted in the
[03:35] study of pragmatics and implicature. Our
[03:37] decomposition tool explores how shared
[03:40] understanding may emerge through subtle
[03:41] and implied meanings. I'll present you
[03:43] the foundation of our decomposition.
[03:46] >> So this is the first part. Uh there's
[03:49] this one here I think that contains
[03:52] either part two or part three. I don't
[03:55] remember. Third phenomenology of int.
[03:57] >> Yeah, that's the third which I remember
[04:00] is needs a a crop. Um
[04:09] anyway the
[04:11] what we have on the left is is not
[04:13] exactly what we want. We we have it at
[04:16] the end.
[04:16] >> Third phenomenology of intentionality.
[04:19] This field studies how consciousness is
[04:20] always directed towards something. It
[04:22] includes noa noises, second order
[04:23] intentionality and temporal
[04:24] constitution. In application, it helps
[04:26] us model intentional state graphs,
[04:28] recursive selfmodel layers, and temporal
[04:29] horizon buffers for reflective agents.
[04:37] So food for thoughts
[04:50] >> third phenomenology of intentionality.
[04:52] This field studies how consciousness is
[04:53] always directed towards something. It
[04:55] includes noa noises, second order
[04:56] intentionality and temporal
[04:57] constitution. In application, it helps
[04:59] us model intentional state graphs,
[05:01] recursive selfmodel layers, and temporal
[05:02] horizon buffers reflective agents.
[05:06] >> That's going to be interesting to have
[05:08] this uh Mietta uh part three of episode
[05:13] 44
[05:15] extracted and obviously probably
[05:18] reddrafted. Uh but there really is um
[05:22] it's an interesting aspect this third
[05:24] one. Uh if we go over it again uh
[05:28] separating the aspect
[05:31] >> third phenomenology of intentionality
[05:35] >> so there's an academic field that study
[05:38] intentions
[05:39] and this is the one
[05:42] >> this field studies how consciousness is
[05:44] always directed towards something.
[05:46] >> So that's an interesting aspect.
[05:48] Consciousness is always
[05:51] directed towards something.
[05:55] >> This is always directed towards
[05:56] something.
[05:58] >> So it aligns with uh the creative
[06:00] orientation
[06:02] uh where there are two modalities. You
[06:05] your problem solver
[06:08] you know there's a fire you need to
[06:10] extinguish
[06:12] and every gap analysis uh will be useful
[06:16] for
[06:18] getting rid of something as fast as we
[06:20] can. uh and there's the other
[06:21] orientation where uh I think this might
[06:25] be where consciousness uh manifests
[06:27] itself which is when it's time to create
[06:32] therefore
[06:34] it would support here uh at the level of
[06:38] what consciousness is and there really
[06:39] is a an interesting word uh
[06:44] in inu
[06:45] uh speech recognition is not going to
[06:48] understand it but it it says mamu
[06:52] uh and and it explores uh how you change
[06:56] consciousness with others
[06:59] and you don't change consciousness
[07:01] obviously as you change a piece of clo.
[07:04] So linear language uh are limited here
[07:07] in expressing and compressing this this
[07:10] uh interesting aspect of relating with
[07:13] uh the creative force uh which is the
[07:16] other world uh that exists. So we can
[07:18] probably find a grounding on um
[07:21] literature uh for indigenous studies
[07:23] here that supports this I'm going to be
[07:27] drawn by consciousness. Uh it means not
[07:30] my ego or my my self-destructive part or
[07:34] sick part or you know this this little
[07:36] uh individualistic
[07:38] uh part of my being that uh maybe it's
[07:42] not even part of my being. It's just
[07:44] like something that uh
[07:47] is stuck here and deny and is colonized
[07:50] basically denying or refusing or even
[07:54] like the DSM1 considers people that uh
[07:57] speak with the great spirit as uh as
[08:01] being schizophrenic. Uh obviously new uh
[08:05] DSM standard uh consider that it's okay
[08:09] to have a spiritual connection. Um
[08:12] obviously it refers to all of the
[08:15] tremendous success of 12step program
[08:17] which uh are really based on the fact
[08:19] that there's a that power greater than
[08:22] yourself that's going to
[08:25] have good inside and it's going to have
[08:28] loving presence for you and guiding your
[08:30] life which normal people uh actually um
[08:34] and I said normal people I said people
[08:36] that are really constructive and where
[08:38] their consciousness is always driven
[08:41] toward for it u
[08:43] something that's constructive, positive
[08:46] and and optimistic and and they're going
[08:49] to be really good contributors to their
[08:51] own community uh great self-care and all
[08:54] these things. So on this loop here uh we
[08:58] have a great sets of elements for uh
[09:01] putting in the review as well as uh it's
[09:04] going to be a good draft for uh my
[09:06] potential contribution. So we're going
[09:08] to continue this uh we are looking at
[09:11] this lesson as a potential uh tool but
[09:15] more than just a tool to obviously give
[09:19] a direction to our consciousness or have
[09:21] a tool to observe the direction that we
[09:23] take with our consciousness
[09:25] >> includes noa nois second order
[09:27] intentionality and
[09:29] >> so there's noa noasis so it's going to
[09:32] be interesting to reexplore and revisit
[09:34] that
[09:38] >> the second order intentionality and
[09:39] temporal constit.
[09:40] >> So the second order intentionality uh
[09:45] in my work with Robert Fritz I think we
[09:48] talked about what is implied uh
[09:51] implicitly.
[09:54] So that's an implicit questioning uh if
[09:57] we think in pictures it's what's in the
[10:00] negative space.
[10:02] I think
[10:06] >> includes noa noises, second order
[10:07] intentionality and temporal
[10:08] constitution.
[10:10] >> So temporal constitution, I have no idea
[10:13] what that is. The only thing that comes
[10:15] to my mind is uh this uh timeline where
[10:19] you constitute something about
[10:22] the decomposition of your
[10:24] intentionality. So basically um what we
[10:27] do we look at the implication of things
[10:29] we try to create a dependency of actions
[10:31] in between them uh because obviously we
[10:34] want to direct the consciousness to
[10:35] achieve results but constructively and
[10:39] really uh building up something
[10:40] interesting. So there might be an
[10:42] element here uh that will help us
[10:46] resolve that aspect
[10:48] >> in application. It helps us model
[10:49] intentional state graphs for
[10:52] >> so basically I will be interested in
[10:54] really making sure that we use
[10:57] an intentional state graph and or that
[10:59] we use at the uh core level of our data
[11:02] structure and engine uh something that
[11:05] supports uh state graph and it might uh
[11:09] help us create the first dependency of
[11:12] this prototype with the actual medicine
[11:15] wheel uh because we really try to
[11:19] Find a home that is uh
[11:22] indigenous in nature. Here
[11:25] >> self model layers and temporal horizon
[11:27] state graphs recursive selfmodel layers
[11:29] and
[11:30] >> recursive selfmodel layers.
[11:35] So it's gonna I think we we support that
[11:37] in a way that uh it it will need
[11:40] revision but there are multiple sessions
[11:43] that starts inside of the ceremony that
[11:47] is used by this talking circle to
[11:49] actually decompose everything. uh and I
[11:53] think that this is where we do a certain
[11:55] amount of recursion
[11:58] >> self model layers and temporal horizon
[12:00] buffers reflective agents
[12:02] >> temporal buffers reflective agent
[12:08] okay that's that's a good startup for
[12:10] the system instructions you are a
[12:13] temporal
[12:15] I don't even remember
[12:17] >> and temporal horizon buffers reflective
[12:19] agents
[12:20] >> you are temporal reflection
[12:23] agent or something like that.
[12:26] So if I go back to this um
[12:31] this here um
[12:34] there's going to be we are back on the
[12:36] uh user interface screenshot but
[12:39] obviously we will want to uh
[12:42] uh be very interested in in how we can
[12:45] have a discussion with the memory about
[12:47] uh the prompt decomposition engine is is
[12:51] actually a a folder where we store all
[12:53] the data uh that is being explored at
[12:57] different uh ways of organizing the
[12:59] milestone. The issues that are grouped
[13:01] inside of it and the actual um talking
[13:05] circle that are going to accompany uh
[13:08] these
[13:09] agent sessions uh to reach their mission
[13:12] of uh supporting basically complex
[13:15] problem that are impossible to
[13:17] comprehend with linear thinking.
[13:22] So on this uh we would have um an
[13:25] interesting uh
[13:28] review of these uh these episode 44 and
[13:33] basically relates to what we talked
[13:35] about today. This is going to be
[13:37] interesting to uh to advance this little
[13:40] episode here
[13:44] which originally uh there's other
[13:46] chapter that we never worked on and I
[13:49] think that this is her minoric
[13:53] and
[13:56] subbernetic science and the theory of
[13:59] mind. So it's going to be interesting.
[14:01] It's going to be a potential followup to
[14:03] uh what we uh explored in these. I think
[14:08] we have three
[14:10] main element to decompose
[14:13] and structure accordingly. If we want to
[14:15] do some work uh on the systems which
[14:18] working on the production of this
[14:20] episode 44 and its academic uh
[14:23] constitution, it's going to
[14:26] require us to work on my
[14:30] creative desk where I'm going to have to
[14:33] write and work on this as well as uh
[14:36] obviously
[14:39] using the whole content that we talked
[14:40] about while reviewing these video
[14:43] because it's going to influence some
[14:44] implementation and some validation some
[14:46] enhancement on the actual uh tool and
[14:50] services uh that that tries to uh
[14:55] put into application this and
[14:59] basically The the other aspect is is
[15:01] going through what the agent will have
[15:04] done in the terminal. So
[15:08] I said that we might want to look at uh
[15:11] how I prompted it u which is going to be
[15:14] a good way to conclude here. Uh if I go
[15:17] back at the beginning it was interesting
[15:19] um that I knew I had a certain amount of
[15:23] iterations that I would give. So my
[15:26] basic uh first prompt was as simple as
[15:29] um we're planning enhancement for the
[15:34] user interface. Uh let's uh look at
[15:40] how Mia and Ava would interpret that.
[15:50] I'm working on app experience
[15:52] enhancements. I'll push changes to main
[15:54] and back into production using a cheaper
[15:56] sub aent model to validate and design
[15:59] the interface. I'll rely on Jev to
[16:01] decide related upgrades, update
[16:03] packages, and plan a release so the
[16:06] community benefits so your leading
[16:08] interface design m
[16:11] agent approach and the package upgrades
[16:14] to keep validation cost effective and
[16:16] ensure the cycle closes cleanly. I'll
[16:19] wait for five screenshot messages
[16:21] guiding five enhancement areas before
[16:23] final deployment. That's a structured
[16:26] rhythm for user experience work. Why
[16:29] wait for exactly five visuals? Each
[16:32] screenshot steers one enhancement area.
[16:35] After five, I publish and close the
[16:37] loop. Your workflow keeps the design
[16:40] domain tight and sharable.
[16:44] >> That's perfect. Really? Uh and you see I
[16:48] didn't gave it the five
[16:53] screenshots. So, it puts the agent in a
[16:55] loop. I'm expecting five messages with
[16:59] screenshot and directions for user
[17:00] enhancement.
[17:02] Uh the second part is says I'm
[17:03] interested in what she understood
[17:05] because uh there was this uh decision
[17:07] engine episode
[17:10] 564 that might have been put into
[17:13] application. Uh but it's going to be
[17:16] needed to enforce that at the harness
[17:20] level of the plug-in which is going to
[17:22] be the aspect number four that might be
[17:25] discussed here.
[17:34] I'm working on deployment automation for
[17:36] GIA. The main branch is synced and colon
[17:40] 3,335
[17:42] is running a production build. Our app
[17:44] myod packages are workspace linked. So
[17:47] any upgrade means publishing and bumping
[17:49] versions. Another seat has uncommitted
[17:52] edits, but I'll only commit my own
[17:54] lines. Unttracked files under app and
[17:56] packages will ship like a spec doc from
[17:59] August. So I'll list what goes into each
[18:01] build. What's the workflow you're
[18:03] orchestrating MIA? Each screenshot
[18:06] triggers five steps. I map roots and
[18:09] components into coded cards. Score them
[18:11] with jev. Anything marked would do now
[18:14] goes live this pass. Then switch colon
[18:17] 3,335
[18:18] to dev mode. A sonet subent checks
[18:21] design. A hiq subagent checks the live
[18:24] page. I commit, push, publish touched
[18:27] packages, rebuild production and
[18:30] restart. Finally, I report changes with
[18:32] a live screenshot. Why this structure in
[18:35] your release cycle? Every commit
[18:38] references an issue. I'll open one in
[18:40] drill mad for the cycle unless it
[18:43] exists. After five areas are live, I run
[18:46] the delivery script and send a summary.
[18:48] Each screenshot becomes a visible change
[18:50] on colon 3,335
[18:53] before the next one. So your release
[18:56] pipeline keeps feedback tight between
[18:58] builds.
[19:04] >> So we see on the screen the uh what we
[19:07] had before and what I gave the agent.
[19:09] I'm not going to go over the prompts,
[19:10] but uh that was uh interesting to uh
[19:16] could be abstracted from the actual uh
[19:20] so it's the wall API.
[19:24] So they they started working on it
[19:27] giving screenshot and and obviously I
[19:29] think the important element of my
[19:31] prompts is always reusability and having
[19:35] something that module well and is that
[19:39] you can use elsewhere.
[19:42] Uh there's one important thing is uh
[19:45] this here is is a complex prompt because
[19:48] I'm planning out the inside of the user
[19:50] interface and I'm using another
[19:52] prototype uh that actually has a really
[19:55] great interactivity for when we go on
[19:58] the land uh and
[20:01] I want that to be used in there but um
[20:05] and also in my author's desk. Uh
[20:09] therefore the prompt was uh complex
[20:11] enough to try to reach that aspect and I
[20:15] had to um
[20:18] correct the agent once
[20:20] because it was starting to do production
[20:22] build and all of these things. So I had
[20:24] to tell him I'm not done with these five
[20:27] message so don't start to build and
[20:29] everything. and it finished a review.
[20:33] And um
[20:35] it received uh obviously the um this
[20:39] last prompt here was uh uh complex
[20:44] enough because it touched uh the whole
[20:46] aspect of research, integrating
[20:49] inquiries, drafting u what inquiries
[20:52] might become when uh doing public
[20:56] facing. Uh
[20:59] so basically uh I don't know if I
[21:01] captured some uh some screenshot of
[21:04] that. I don't think so.
[21:07] It was empty anyway. So they worked uh
[21:09] you see that um they got down the stack
[21:12] to actually um
[21:15] go down to the medicine wheel memory uh
[21:17] to try to uh prepare some wrapping
[21:20] prompts cuz obviously if you want to
[21:21] question the memory from the homepage
[21:23] it's not going to be the same uh
[21:25] prompting uh that if we question the
[21:27] memory from uh
[21:31] within your talking circle. Also, I'm uh
[21:33] trying to have some uh prompting prompt
[21:36] engineering layers uh that we'll be able
[21:39] to uh embed some variables from the the
[21:43] user input from the wall. So we see that
[21:46] they developed uh at the lower level a
[21:50] new version of the medicine wheel to
[21:52] actually um
[21:55] support that basically and and obviously
[21:58] hopefully put uh this contribution
[22:00] available to the community
[22:03] or at least put the code at the lowest
[22:06] level.
[22:09] So I this one is my last prompt. Yeah.
[22:13] So it touches the um there really is so
[22:16] it's the it's a fifth aspect somehow
[22:19] here of this recording because
[22:23] and it relates to everything else
[22:25] because uh custom GP2 going to be
[22:27] deprecated in December. So I'm having
[22:30] this plugin being uh developed uh and
[22:33] this is an important element here.
[22:35] There's this book really important book
[22:38] it teaches um
[22:41] how to do research ingestion
[22:44] with respect I should say it's going to
[22:46] be a value here that and and this
[22:49] literature
[22:51] uh would influence how we we uh work
[22:54] with data basically uh so this agent
[22:58] here has done this uh really amazing
[23:00] deep research with a nice prompt and
[23:04] they produced used
[23:06] this markdown and usually I download it,
[23:09] I put it in the folder, I create an
[23:10] issue inside of the inquiry uh
[23:13] repository that is meant to be shared by
[23:16] various communities. Uh and then I
[23:18] commit, I push, I relate that to my
[23:21] internal systems for for talking about
[23:23] it inside of the internal systems to try
[23:25] to see how I'm going to use it. uh what
[23:27] are the different questions and try to
[23:29] basically apply it to my own process.
[23:32] And it's it's okay to do it, but there
[23:35] really is a need for delivering that
[23:37] with agents because there's going to be
[23:39] a um chains of of agent that are going
[23:43] to see what inquiries the community
[23:46] might be interested on so they can uh
[23:49] fork it or vote on it or or try to uh
[23:54] not approve it in a sense of uh
[23:55] complacency but uh supporting it
[23:58] somehow. Uh so there's this uh
[24:03] workflow that the MCP tools goes through
[24:06] uh to be able to uh create the inquiry
[24:09] artifact uh and actually uh
[24:13] I can approve it uh I can approve with
[24:16] changes I give instructions
[24:18] uh or I send back to revision
[24:21] with this payload here. So obviously I
[24:24] had to tell the agent uh because it was
[24:27] um about to work on my research page for
[24:32] user interface
[24:33] uh and it was a complex uh request here.
[24:37] Uh and I think that that might be where
[24:39] the agents actually uh uh drifted. I'm
[24:42] not sure. So what I did on the computer
[24:44] is I I pasted the actual payload uh and
[24:47] told the agent that uh it had to work um
[24:51] on these. Remember that I waited a
[24:53] little bit to actually uh tell them I'm
[24:56] going to approve that request and I want
[24:58] you to work on that. You're going to
[25:02] make sure that all the services layer
[25:04] below are supporting uh this this actual
[25:08] work. And because it's it's the second
[25:11] or third time that I go over revising uh
[25:14] this plugin
[25:17] uh obviously doesn't deliver what it
[25:19] should deliver. So basically the slash
[25:22] research should be uh your own personal
[25:24] space where you're capable of managing
[25:26] your personal inquiries, your collective
[25:29] deep research to expose them to the
[25:31] community so they can uh do something
[25:33] about it which is an area where you need
[25:36] feedback and potential idea of how you
[25:39] would as a member of the community with
[25:42] different sets of permission use it in
[25:46] your applicable domain. And uh so yeah
[25:50] so basically this is a little where uh I
[25:53] would do this approval that we saw
[25:55] before not necessarily just from uh what
[25:58] was interesting in here.
[26:01] Um so this is where I enter my token is
[26:04] I don't have to quit
[26:06] chat GPT. So we see here I'm having my
[26:10] discussion and uh as soon as I click
[26:13] open uh to show the request that always
[26:16] decision there's an embedded system that
[26:19] ask about my token and then shows uh the
[26:21] actual um uh form in here. So I don't
[26:25] have to uh
[26:28] to go over a different user interface.
[26:30] So obviously this uh
[26:34] I should say that to develop an inquiry
[26:36] here should lead us to have a an
[26:39] enhanced user experience.
[26:42] Pasting tokens all the time is not
[26:44] necessarily
[26:46] awesome and also would be interested to
[26:50] work on a on a conversation and chat
[26:53] GPT. It could be someone else in the
[26:55] community that has a responsibility and
[26:58] there's an inquiry that gets submitted
[27:00] and
[27:02] I will see it in my uh slash research.
[27:06] Uh so after that there's a bunch of
[27:09] words that I have no idea uh what are
[27:11] the actual uh final results.
[27:14] uh they produce a little bit of um
[27:19] you see we we're going to have uh on on
[27:22] the top right a reusable component to
[27:24] ask. We're back on the first aspect here
[27:27] which uh was the user interface
[27:29] enhancement and everything but obviously
[27:31] that relates to everything else. You're
[27:33] going to be capable of asking the
[27:35] circle's memory from uh the walls. Uh
[27:39] so this is an example of the output uh
[27:42] that it gives. Obviously it's meant to
[27:45] give you insight but also helps you uh
[27:48] we talked about intentionality
[27:50] uh in consciousness before for episode
[27:53] 44. Uh and and this is where you might
[27:56] ask the circle's memory to be directed
[27:59] toward what what you want. And obviously
[28:02] sometimes it's not the first cycle that
[28:04] you do. So you you want to be directed
[28:06] to the the adequate talking circle and
[28:08] ceremony that exist or a place where you
[28:11] can open a new ceremony in an existing
[28:13] talking circle or just know and have all
[28:16] these reference like you select
[28:18] everything else there you paste that
[28:20] agent and you say we're going to relate
[28:22] to that. Here's what I want to do.
[28:25] Here's what I want to talk about.
[28:26] Something like that.
[28:28] Um
[28:30] so obviously I had the agent uh create
[28:32] the search center uh which basically is
[28:35] going to uh uh these are whole uh ways
[28:38] to actually work uh that we had in the
[28:41] first version where you have your
[28:43] redstone and your lattice. So I had that
[28:46] uh removed from the user interface
[28:48] homepage where which was just a little
[28:50] confusing and obviously it's going to
[28:52] try to uh
[28:55] consolidate and unify everything. So you
[28:57] should be capable of searching uh the
[29:00] whole factory uh which includes uh the
[29:04] chronicle episode probably the redstone
[29:06] and so and u the memory as well. I have
[29:10] no idea what's going to look like uh and
[29:14] we see that it tries to relate uh to my
[29:17] research. Oh okay there's more
[29:19] information I'm discovering at the same
[29:21] time here uh the actual result. So it's
[29:24] going to look in the circles in the
[29:26] inquiry delivered report and request.
[29:29] Interesting.
[29:30] Uh and it's going to basically it it
[29:33] seems like the whole uh unification of
[29:36] the search is uh really in good
[29:38] progress.
[29:40] Uh so this is an example report. Uh so
[29:43] that's interesting.
[29:46] So we see here that what I delivered
[29:48] from the other with the other agents
[29:50] like report delivered uh with the
[29:54] literature and the research network. Uh
[29:59] research network is I um always uh ask
[30:03] for a picture a graph of the
[30:05] relationship of a an author when I do
[30:08] the literature review. So I know with
[30:11] who else they are related uh and pretty
[30:14] quick within the engineering systems
[30:16] there's always uh a relative u that I
[30:20] worked with its knowledge before so as
[30:23] Sean Wilson would say I'm in a
[30:25] relationship with is book researches
[30:27] ceremony therefore anybody else that is
[30:32] related and in relationship with it or
[30:34] that Shauna Wilson was in relationship
[30:36] with in his biography obviously
[30:40] is in relationship with someone else
[30:42] which I might be able to create a
[30:45] relationship with her.
[30:48] So basically we all we even see the
[30:50] failure to approve and everything that's
[30:52] probably going to get fixed. And there's
[30:54] also obviously this um Jev engine
[30:58] evaluation gate which uh comes from the
[31:01] chronicle 564 where there is a big
[31:04] exploration of uh what's going to happen
[31:08] with with decision everything. So I
[31:10] found found out some result about that.
[31:12] It's going to be pretty interesting.
[31:16] So that is not much.
[31:21] This is the view of uh the ceremony in
[31:25] progress. So probably they did some uh
[31:28] some enhancement there as well.
[32:04] So here we see that they they got the
[32:06] search center better obviously because
[32:09] I'm trying to create this recursive loop
[32:12] with the decision engine in the review
[32:14] process with sub agent.
[32:16] Uh it seems like they decided to go into
[32:19] one more recursion and actually decide
[32:23] oh yeah the user going to be able to
[32:26] desired recording and transcribing
[32:28] themsel and not typing what they want to
[32:31] search for.
[32:33] Uh therefore it's a good decision here
[32:37] that they took
[32:42] and after that they seem to have worked
[32:44] on revising the connector because
[32:46] obviously there was some issues with the
[32:48] um
[32:50] the approval that we saw in previous
[32:52] transcript. So they probably worked on
[32:54] on that uh as well.
[33:05] So we see here they refine the search
[33:07] center.
[33:10] So you have these uh reusable component
[33:13] that searched and present. There's are
[33:17] the result in your research. Uh I don't
[33:20] know how to call it yet but uh it's
[33:22] basically where you manage your
[33:23] inquiries.
[33:26] That would be where
[33:28] the cycle circle memory would result
[33:31] would be because we've never talked
[33:33] about her research on these inquiries.
[33:35] That's probably why there's nothing
[33:36] there.
[33:38] That's going to be amazing.
[33:53] So, it's going to be a prompt template
[33:55] called circle memory ask. See when we
[33:58] talked about that before.
[34:00] It's a medicine wheel 171.
[34:09] And if we go over the the it sounds to
[34:12] be an important output. Um, area one is
[34:17] where you sit.
[34:19] Area two is a circle memory and
[34:21] homepage. Area three is the record and
[34:24] hear yourself and transcribe everything.
[34:28] Area four is the search center and area
[34:31] five is the slash research
[34:35] to manage basically slash researches uh
[34:38] for doing researches ceremony I guess
[34:44] in academic way I should say.
[34:48] Uh so there's a bunch of other things uh
[34:52] that obviously waits on my words.
[34:57] So let's finish this screenwalk on that
[35:00] gives feedback on what's waiting for my
[35:02] words here.
[35:06] Waiting on your word Jeff put none of
[35:08] the related cards above its 0.8 line. So
[35:11] I built only what you asked. JG will
[35:14] my84 lists the rest. the ones that need
[35:18] a decision from you. E8, a package name
[35:21] for the slice component. E36, renaming
[35:24] the nav's memory entry, no C, and
[35:27] pointing it at search. E28, the desk
[35:30] companion that sorts a spoken take into
[35:32] item types. My take, a question, a
[35:35] spark. The Quimless request is still
[35:37] approved. Run failed. The cause is
[35:40] fixed. And run it again on research.
[35:42] We'll run it with your approval and
[35:44] instructions. I left that click to you.
[35:46] You can now ask what all your circles
[35:48] remember, say a turn instead of typing
[35:50] it, and run a failed approval again
[35:53] yourself without waiting on an agent.
[35:58] >> All right.