December 5, 2025 • 15 mins
For the first time ever, Jennifer welcomes a guest to the show! William Lyon gives us a deep dive into the evolving world of AI agents, knowledge graphs, and the concept of memory in artificial intelligence.
Episode highlights:
- William’s career journey: from Neo4j to startups and back again
- The role of knowledge graphs in agentic memory and reasoning
- Types of memory in AI agents: episodic, procedural, and more
- How knowledge graphs can model both user-facing and operational memory
- The importance of domain-specific data modeling for AI memory systems
- William’s AI Memory Landscape project: cataloging tools, frameworks, and services in the AI agent memory space
- Contributions to the project are open, so submit a PR or request!
- Advice for developers architecting AI agents with memory
Other referenced links:
- GraphStuff.FM podcast
- AI Memory Landscape project: https://ai-memory-landscape.netlify.app/
Connect with William Lyon:
- Website: https://lyonwj.com/
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Jennifer Reif (00:05):
You are listening to the Breaktime Tech Talks podcast, a bite-sized
tech podcast for busy developers wherewe'll briefly cover technical topics, new
snippets, and more in short time blocks.
I'm your host, Jennifer Reif, anavid developer and problem solver
with special interest in data,learning, and all things technology.
Welcome back today we are gonna dosomething a little bit different.
(00:28):
So I have a guest on the podcast thatI'm gonna talk to, and we're just
gonna discuss either what's going on,what the person's working on projects,
things in the industry, and so on.
I am really excited to welcomeWilliam Lyon to the podcast.
Hi Will!
William Lyon (00:45):
Hello, hello.
Jennifer Reif (00:47):
So Will used to work for Neo4j and then took a little bit
of a sabbatical explored some otherthings, other projects, other companies.
And then has come full circleback to Neo4j recently.
I am really excited to kind of hearhow that journey went and the things
that he explored and did along the way
William Lyon (01:06):
That's right.
Yeah.
I, I think I am less than a monthback at Neo4j, this time in a role
on, on the product team, right?
I, I worked previously with, withJen and the team on Developer
Relations for, for a long time.
And then, yeah.
Jennifer Reif (01:22):
If you listened to GraphStuff.FM that was all
Will's pretty much idea, I think.
And he, he ran and hosted thatshow for the longest time.
William Lyon (01:32):
That's right.
Yeah.
I, I love that podcast.
Is, is it gonna get rebooted?
Is there, is there an encorecall for GraphStuff.FM?
Jennifer Reif (01:40):
Has been discussed and surfaced occasionally, but
I don't have any like insidernews to provide right now.
William Lyon (01:49):
Noted.
Jennifer Reif (01:51):
So would you like to just talk about maybe some of the projects
you've been working on recently orthings that you did between Neo4j stints?
I know you were working on
William Lyon (02:03):
Yeah.
Well,
Jennifer Reif (02:04):
a lot of AI things.
William Lyon (02:06):
Yeah.
E, e, exactly.
Yeah.
So I, I was away from Neo4jfor a little over two years.
I did developer relationsat an early startup.
I think we were eightpeople when I joined.
So I was building out thedeveloper relations function there.
It was a geospatial analytics startup,which was, was really interesting.
(02:28):
Just ended up not beinga great fit for me.
I dunno, technology ecosystem wise, whichI hadn't really appreciated, I guess, how
important I think being passionate aboutthe, the technology and the technology
ecosystem is for develop relations.
So that was an interestinglearning experience for me.
And then I went to go work at anotherstartup called Hypermode, which
(02:50):
is doing really interesting thingswith AI agents and knowledge graphs.
So I got got back in thegraph game there doing really
interesting things with AI as well.
And then, yeah, that, that startupfolded into some other things,
and now I'm back at Neo4j on theproduct team this time looking
specifically at AI innovation projects.
(03:13):
So trying to take some of the, thelearning and observations that I
had working in the, the AI startupecosystem for a little bit and pair
that with what folks are doing withNeo4j using knowledge graphs for,
for GraphRAG, these types of things.
What are other interesting things thatwe could maybe build either into Neo4j
(03:37):
or maybe build on top of Neo4j to supportdevelopers who want to build with LLMs.
Right now, this is mostly, I, I think,agents, lots of interest in, being
able to put LLM and AI agents intoproduction and, and graphs have several
interesting roles to play there.
So that's what, what I'mdipping my feet into now.
Jennifer Reif (03:58):
Okay I was gonna ask what were what are kind of your
topics that you see as being reallykey in the industry right now.
I know you mentioned agents andknowledge graphs, but are there any
other things that you're kind ofexploring those avenues as well?
To integrate?
William Lyon (04:17):
Yeah, so I think that, for folks building AI agents, using, using
knowledge graphs is really importantfor enabling things like agentic memory.
So we see this, this memory functionalityin various agents, but there are
(04:37):
actually some really interesting thingsyou can do as you are constructing
a knowledge graph from interactionswith the agent and the user.
And you're using, that knowledge graphto build up the agent's memory over
time, that can really improve, you know,the, the interactions with the agent.
(04:57):
This is typically, what'scalled episodic memory.
But there, there are othertypes of memory that, that are
typically thought of, as well.
Things like procedural memory,which is an area that I'm
dipping into now, looking at how.
You know, modeling the reasoning,the, the tool calling steps that go
(05:22):
on during an agentic loop, and usinga graph to keep track of the different
execution paths, the tool calls the,you know, like number of tokens.
You also think of this as, you know,model traceability, these sorts of things.
But this, this is verymuch a graph structure.
So if we are using Neo4j to build outthis like procedural memory graph, how
(05:45):
can we then leverage that to supportthe reasoning phase for the agent,
sort of like graph based reasoning?
So that, that's an area that I think is,is out there., There, there's been some
research in that area, but I, I thinkwe're going to see some really interesting
implementations of that pretty soon.
So that's another exciting area.
Jennifer Reif (06:04):
I find it really interesting actually that
both of the memory pathways, Iguess, that you talked about.
One was a very real time conversation,here's what's going on right now in the
conversation and here's all the previouscontext from this existing conversation.
And then the other side of it isactually the as you said traceability,
(06:28):
the operational end of things.
Like what has happened over timeand what is the process underneath
the hood that's going on andare there ways to optimize that.
I just found it really interesting thatone is very time almost user focused,
customer facing and the other one isvery operational internal type of work
William Lyon (06:53):
Yeah, that's a good observation.
The, the agentic memory area is reallyinteresting because the academic
literature is very rich in this area.
Drawing on things from neuroscienceand our understanding of, of
cognition and, and how like humanmemory and the human brain works.
(07:13):
And there's a lot of, a lot ofresearch that is taking those
concepts and applying them to sortof the tooling around AI agents.
And, and so, you know, yourobservation that, hey, we have short
term and, and long term memory.
The, the mental model that is sort oftalked about in the literature maps to the
way the human brain works with differencesand short-term memory, long-term memory.
(07:36):
How we have like a model of something,like a model of a coffee cup, right?
Like this, this mug I have.
Just an understanding of, of what a mugis, and I can go to that, like canonical
representation of, of the thing in mybrain, which when I'm reading some of
the, the neuroscience books and tryingto understand that, like is, is really
(07:57):
interesting to me because that's exactlythe benefit of a knowledge graph, right?
You have like a canonical representationof the, the thing, the, the node.
And as you go through yourinteractions with the agent, as, as
you build out the knowledge graph,you're not duplicating every time
you think of the, the coffee cup.
You're using that, that node,that, that model for it.
(08:18):
And that's exactly the same way thatthat sort of memory works in the brain.
And, and so I think that that's just likeone, one example where using a knowledge
graph for, for agent memory can be reallyuseful and, and sort of maps to, yeah,
some of the, the research out there.
Jennifer Reif (08:33):
Yeah that's true.
I guess with graph use cases, the flexiblething about them is you can choose what
you're trying to model and you can modelthe same entity, for instance the coffee
cup or a mug, and you can model itdifferently depending on what you're using
it for, whether that's a random householditem or it's something to hold liquids
(08:57):
or it's something else entirely, right?
William Lyon (09:01):
Totally.
Yeah.
Like graph data modeling, fundamentally,I, I think is one, is like this
iterative process, but also verydependent on the questions that
you're asking of the data, right?
And, and so I think that conceptfundamentally applies also in the
agentic world where we're buildingknowledge graphs for AI memory or graph
(09:27):
reasoning, these sorts of things as well.
If we're able to sort of apply the,the domain that we're interested
in or apply the, the data modelthey're interested in during sort of
the extraction phase of identifyingentities that are relevant for, for
memory during our agent's interaction.
That's gonna give us a muchmore domain specific and, and
(09:49):
relevant knowledge graph, right?
Than if we're just trying tosort of construct a knowledge
graph of every entity that we'vetalked about in, in our memories.
So I think that's also like areally important piece too, to, to
think about as you're building, youknow, sort of, AI memory systems.
Is to, to think about during that unityextraction phase, like what are the
(10:13):
things that I'm talking about in, inthese interactions with, with the agent?
Think about applying a, a domain tothe data model for that and, and only.
I, I guess the, the first step would beonly create entities and extract things
that are relevant for the domain ofwhatever application you're building.
That's a good first step, but ideallyyou're able to use some sort of like
(10:37):
structured output technique to map the,the types that you are constructing to
the types in your domain application,so that you are able to link sort of
the, the canonical representation of thething that, that you're talking about.
If you are talking about youragent, about a coffee mug.
(10:58):
If you're able to link that withyour, I don't know, your e-commerce
inventory internal database.
What, what row in that tablethe coffee mug is, right?
Like, if you're able to link thosetwo things, then I think you're
gonna leverage a lot more valueout of building this knowledge
graph from your agent interactions.
Combining that with internalbusiness data, that's like a big
(11:22):
power of the graph, combining dataand querying across silos, right?
That I, I think, givesa lot of value there.
Jennifer Reif (11:29):
Okay.
Awesome.
I was also gonna ask you'd mentionedit in another conversation that you're
working on an AI memory landscape project.
And I popped out to it, and it's like thisbig dashboard with several categories.
Could you talk a littlebit about the project?
Is it something you've built or is thatsomething that's a wider industry or
vendor something that's been assembled?
William Lyon (11:50):
Yeah.
So, as I mentioned, I'm in, ina relatively new role in, in,
in the product team at Neo4j.
And, I think during, during anynew role like that, you go through
this research phase to try to get,get the lay of the land, understand
the, the landscape of tooling, opensource projects, other vendors.
(12:11):
How the pieces fit togetherI, I, I think is important.
And,
Jennifer Reif (12:13):
Sure
William Lyon (12:14):
I, I thought, well, if I'm doing this for me
to understand what's going on.
The entire AI ecosystemdevelops so rapidly.
There's a lot of interesting researchgoing on just in the, the AI memory space.
I think this is, this is rapidlyevolving too with the like product
offerings that are coming out fromlike some of the model providers,
some of the, the agent framework.
(12:34):
So there's a lot going on,and there's to maintain.
And so my first thought was canI just share and make public my
understanding of this for other folks.
And then, selfishly, secondly, is if Iput something together that's useful,
could I maybe encourage folks tocontribute and, and help us update it?
(12:56):
And so, so I'm at phase one now.
It's still a, a work in progress.
Maybe by the time the podcast airs, it'llbe public, but I, I think you've seen like
the, the early preview version of this.
But yeah, that's really the goal isjust to, to catalog what are the pieces
of the AI agent memory landscape.
(13:18):
If I'm a, a developer, and Ineed to add memory to my agent.
If I'm architecting a newproject from scratch, what,
what models should I choose?
What agent framework?
What memory services?
Do I wanna managed solution?
What are the, the open source tooling?
What are the, the, forinfrastructure database pieces,
that, that have a role in this?
So that's, that's really the goalof the, the AI memory landscape.
(13:40):
we'll, we'll try to link it inthe, the show notes if, if we can.
And, yeah, really curious to just haveany, any feedback or contributions
from, folks in the community too.
If there's a project you're aware of thatwe missed, open a pull request and add it.
There's also a blog there for, I wantto do like tutorial series, you know,
with more, sort of examples of showinghow different pieces fit together.
(14:02):
Just meant to be like a vendor neutralspace of just showing how to use different
tools in the the AI memory ecosystem.
Jennifer Reif (14:09):
Yeah that's super helpful.
I know for me just trying tofigure out what to use and what
all is available out there.
You have to move across tons ofdifferent surface areas to figure
out what's the best thing here.
What do I really want here?
So having one dashboard, if you will,to showcase it all where you can pick
and choose oh this is the right thingor maybe I wanna go down this path
first, I think is really super helpful.
(14:30):
So I'll definitely add that tothe show notes and share that.
I know you have to wrap up and, and headout, but thank you so much for joining me
on the podcast and discussing a little bitabout what you're, what you're working on.
And I can't wait to see a lot of thecontent and things that you work on next.
William Lyon (14:46):
Awesome.
Yeah, thanks for having me onand looking forward to maybe
chatting again in the future.
Jennifer Reif (14:52):
Sounds good.
All right, bye.
You are listening to the Breaktime Tech Talks podcast, a bite-sized
tech podcast for busy developers wherewe'll briefly cover technical topics, new
snippets, and more in short time blocks.
I'm your host, Jennifer Reif, anavid developer and problem solver
with special interest in data,learning, and all things technology.
Welcome back today we are gonna dosomething a little bit different.
(00:28):
So I have a guest on the podcast thatI'm gonna talk to, and we're just
gonna discuss either what's going on,what the person's working on projects,
things in the industry, and so on.
I am really excited to welcomeWilliam Lyon to the podcast.
Hi Will!
William Lyon (00:45):
Hello, hello.
Jennifer Reif (00:47):
So Will used to work for Neo4j and then took a little bit
of a sabbatical explored some otherthings, other projects, other companies.
And then has come full circleback to Neo4j recently.
I am really excited to kind of hearhow that journey went and the things
that he explored and did along the way
William Lyon (01:06):
That's right.
Yeah.
I, I think I am less than a monthback at Neo4j, this time in a role
on, on the product team, right?
I, I worked previously with, withJen and the team on Developer
Relations for, for a long time.
And then, yeah.
Jennifer Reif (01:22):
If you listened to GraphStuff.FM that was all
Will's pretty much idea, I think.
And he, he ran and hosted thatshow for the longest time.
William Lyon (01:32):
That's right.
Yeah.
I, I love that podcast.
Is, is it gonna get rebooted?
Is there, is there an encorecall for GraphStuff.FM?
Jennifer Reif (01:40):
Has been discussed and surfaced occasionally, but
I don't have any like insidernews to provide right now.
William Lyon (01:49):
Noted.
Jennifer Reif (01:51):
So would you like to just talk about maybe some of the projects
you've been working on recently orthings that you did between Neo4j stints?
I know you were working on
William Lyon (02:03):
Yeah.
Well,
Jennifer Reif (02:04):
a lot of AI things.
William Lyon (02:06):
Yeah.
E, e, exactly.
Yeah.
So I, I was away from Neo4jfor a little over two years.
I did developer relationsat an early startup.
I think we were eightpeople when I joined.
So I was building out thedeveloper relations function there.
It was a geospatial analytics startup,which was, was really interesting.
(02:28):
Just ended up not beinga great fit for me.
I dunno, technology ecosystem wise, whichI hadn't really appreciated, I guess, how
important I think being passionate aboutthe, the technology and the technology
ecosystem is for develop relations.
So that was an interestinglearning experience for me.
And then I went to go work at anotherstartup called Hypermode, which
(02:50):
is doing really interesting thingswith AI agents and knowledge graphs.
So I got got back in thegraph game there doing really
interesting things with AI as well.
And then, yeah, that, that startupfolded into some other things,
and now I'm back at Neo4j on theproduct team this time looking
specifically at AI innovation projects.
(03:13):
So trying to take some of the, thelearning and observations that I
had working in the, the AI startupecosystem for a little bit and pair
that with what folks are doing withNeo4j using knowledge graphs for,
for GraphRAG, these types of things.
What are other interesting things thatwe could maybe build either into Neo4j
(03:37):
or maybe build on top of Neo4j to supportdevelopers who want to build with LLMs.
Right now, this is mostly, I, I think,agents, lots of interest in, being
able to put LLM and AI agents intoproduction and, and graphs have several
interesting roles to play there.
So that's what, what I'mdipping my feet into now.
Jennifer Reif (03:58):
Okay I was gonna ask what were what are kind of your
topics that you see as being reallykey in the industry right now.
I know you mentioned agents andknowledge graphs, but are there any
other things that you're kind ofexploring those avenues as well?
To integrate?
William Lyon (04:17):
Yeah, so I think that, for folks building AI agents, using, using
knowledge graphs is really importantfor enabling things like agentic memory.
So we see this, this memory functionalityin various agents, but there are
(04:37):
actually some really interesting thingsyou can do as you are constructing
a knowledge graph from interactionswith the agent and the user.
And you're using, that knowledge graphto build up the agent's memory over
time, that can really improve, you know,the, the interactions with the agent.
(04:57):
This is typically, what'scalled episodic memory.
But there, there are othertypes of memory that, that are
typically thought of, as well.
Things like procedural memory,which is an area that I'm
dipping into now, looking at how.
You know, modeling the reasoning,the, the tool calling steps that go
(05:22):
on during an agentic loop, and usinga graph to keep track of the different
execution paths, the tool calls the,you know, like number of tokens.
You also think of this as, you know,model traceability, these sorts of things.
But this, this is verymuch a graph structure.
So if we are using Neo4j to build outthis like procedural memory graph, how
(05:45):
can we then leverage that to supportthe reasoning phase for the agent,
sort of like graph based reasoning?
So that, that's an area that I think is,is out there., There, there's been some
research in that area, but I, I thinkwe're going to see some really interesting
implementations of that pretty soon.
So that's another exciting area.
Jennifer Reif (06:04):
I find it really interesting actually that
both of the memory pathways, Iguess, that you talked about.
One was a very real time conversation,here's what's going on right now in the
conversation and here's all the previouscontext from this existing conversation.
And then the other side of it isactually the as you said traceability,
(06:28):
the operational end of things.
Like what has happened over timeand what is the process underneath
the hood that's going on andare there ways to optimize that.
I just found it really interesting thatone is very time almost user focused,
customer facing and the other one isvery operational internal type of work
William Lyon (06:53):
Yeah, that's a good observation.
The, the agentic memory area is reallyinteresting because the academic
literature is very rich in this area.
Drawing on things from neuroscienceand our understanding of, of
cognition and, and how like humanmemory and the human brain works.
(07:13):
And there's a lot of, a lot ofresearch that is taking those
concepts and applying them to sortof the tooling around AI agents.
And, and so, you know, yourobservation that, hey, we have short
term and, and long term memory.
The, the mental model that is sort oftalked about in the literature maps to the
way the human brain works with differencesand short-term memory, long-term memory.
(07:36):
How we have like a model of something,like a model of a coffee cup, right?
Like this, this mug I have.
Just an understanding of, of what a mugis, and I can go to that, like canonical
representation of, of the thing in mybrain, which when I'm reading some of
the, the neuroscience books and tryingto understand that, like is, is really
(07:57):
interesting to me because that's exactlythe benefit of a knowledge graph, right?
You have like a canonical representationof the, the thing, the, the node.
And as you go through yourinteractions with the agent, as, as
you build out the knowledge graph,you're not duplicating every time
you think of the, the coffee cup.
You're using that, that node,that, that model for it.
(08:18):
And that's exactly the same way thatthat sort of memory works in the brain.
And, and so I think that that's just likeone, one example where using a knowledge
graph for, for agent memory can be reallyuseful and, and sort of maps to, yeah,
some of the, the research out there.
Jennifer Reif (08:33):
Yeah that's true.
I guess with graph use cases, the flexiblething about them is you can choose what
you're trying to model and you can modelthe same entity, for instance the coffee
cup or a mug, and you can model itdifferently depending on what you're using
it for, whether that's a random householditem or it's something to hold liquids
(08:57):
or it's something else entirely, right?
William Lyon (09:01):
Totally.
Yeah.
Like graph data modeling, fundamentally,I, I think is one, is like this
iterative process, but also verydependent on the questions that
you're asking of the data, right?
And, and so I think that conceptfundamentally applies also in the
agentic world where we're buildingknowledge graphs for AI memory or graph
(09:27):
reasoning, these sorts of things as well.
If we're able to sort of apply the,the domain that we're interested
in or apply the, the data modelthey're interested in during sort of
the extraction phase of identifyingentities that are relevant for, for
memory during our agent's interaction.
That's gonna give us a muchmore domain specific and, and
(09:49):
relevant knowledge graph, right?
Than if we're just trying tosort of construct a knowledge
graph of every entity that we'vetalked about in, in our memories.
So I think that's also like areally important piece too, to, to
think about as you're building, youknow, sort of, AI memory systems.
Is to, to think about during that unityextraction phase, like what are the
(10:13):
things that I'm talking about in, inthese interactions with, with the agent?
Think about applying a, a domain tothe data model for that and, and only.
I, I guess the, the first step would beonly create entities and extract things
that are relevant for the domain ofwhatever application you're building.
That's a good first step, but ideallyyou're able to use some sort of like
(10:37):
structured output technique to map the,the types that you are constructing to
the types in your domain application,so that you are able to link sort of
the, the canonical representation of thething that, that you're talking about.
If you are talking about youragent, about a coffee mug.
(10:58):
If you're able to link that withyour, I don't know, your e-commerce
inventory internal database.
What, what row in that tablethe coffee mug is, right?
Like, if you're able to link thosetwo things, then I think you're
gonna leverage a lot more valueout of building this knowledge
graph from your agent interactions.
Combining that with internalbusiness data, that's like a big
(11:22):
power of the graph, combining dataand querying across silos, right?
That I, I think, givesa lot of value there.
Jennifer Reif (11:29):
Okay.
Awesome.
I was also gonna ask you'd mentionedit in another conversation that you're
working on an AI memory landscape project.
And I popped out to it, and it's like thisbig dashboard with several categories.
Could you talk a littlebit about the project?
Is it something you've built or is thatsomething that's a wider industry or
vendor something that's been assembled?
William Lyon (11:50):
Yeah.
So, as I mentioned, I'm in, ina relatively new role in, in,
in the product team at Neo4j.
And, I think during, during anynew role like that, you go through
this research phase to try to get,get the lay of the land, understand
the, the landscape of tooling, opensource projects, other vendors.
(12:11):
How the pieces fit togetherI, I, I think is important.
And,
Jennifer Reif (12:13):
Sure
William Lyon (12:14):
I, I thought, well, if I'm doing this for me
to understand what's going on.
The entire AI ecosystemdevelops so rapidly.
There's a lot of interesting researchgoing on just in the, the AI memory space.
I think this is, this is rapidlyevolving too with the like product
offerings that are coming out fromlike some of the model providers,
some of the, the agent framework.
(12:34):
So there's a lot going on,and there's to maintain.
And so my first thought was canI just share and make public my
understanding of this for other folks.
And then, selfishly, secondly, is if Iput something together that's useful,
could I maybe encourage folks tocontribute and, and help us update it?
(12:56):
And so, so I'm at phase one now.
It's still a, a work in progress.
Maybe by the time the podcast airs, it'llbe public, but I, I think you've seen like
the, the early preview version of this.
But yeah, that's really the goal isjust to, to catalog what are the pieces
of the AI agent memory landscape.
(13:18):
If I'm a, a developer, and Ineed to add memory to my agent.
If I'm architecting a newproject from scratch, what,
what models should I choose?
What agent framework?
What memory services?
Do I wanna managed solution?
What are the, the open source tooling?
What are the, the, forinfrastructure database pieces,
that, that have a role in this?
So that's, that's really the goalof the, the AI memory landscape.
(13:40):
we'll, we'll try to link it inthe, the show notes if, if we can.
And, yeah, really curious to just haveany, any feedback or contributions
from, folks in the community too.
If there's a project you're aware of thatwe missed, open a pull request and add it.
There's also a blog there for, I wantto do like tutorial series, you know,
with more, sort of examples of showinghow different pieces fit together.
(14:02):
Just meant to be like a vendor neutralspace of just showing how to use different
tools in the the AI memory ecosystem.
Jennifer Reif (14:09):
Yeah that's super helpful.
I know for me just trying tofigure out what to use and what
all is available out there.
You have to move across tons ofdifferent surface areas to figure
out what's the best thing here.
What do I really want here?
So having one dashboard, if you will,to showcase it all where you can pick
and choose oh this is the right thingor maybe I wanna go down this path
first, I think is really super helpful.
(14:30):
So I'll definitely add that tothe show notes and share that.
I know you have to wrap up and, and headout, but thank you so much for joining me
on the podcast and discussing a little bitabout what you're, what you're working on.
And I can't wait to see a lot of thecontent and things that you work on next.
William Lyon (14:46):
Awesome.
Yeah, thanks for having me onand looking forward to maybe
chatting again in the future.
Jennifer Reif (14:52):
Sounds good.
All right, bye.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!