Slides: Turn 10,994 Notes Into Memory - Paul Iusztin, Decoding AI & Louis-François Bouchard, Towards AI
Source Video
Turn 10,994 Notes Into Memory - Paul Iusztin, Decoding AI & Louis-François Bouchard, Towards AI
Relationship To World's Fair 2026
These slides are extracted from a public AI Engineer YouTube video connected to World's Fair 2026. Speaker-matched clips are supporting context unless later confirmed as exact session recordings; official livestream recordings are day-level/event-level source material.
Related Scheduled Sessions
- No individual scheduled session mapping has been assigned yet; treat this as an event livestream deck.
Extracted Slides
OCR text:
My Second Brain -
5,489 notes in Obsidian aaa Fr 1
rh \
5,505 files in Readwise ie — |
Plus, Notion, Google Drive. ona
=
Growing ~250 files per month = a
.
OCR text:
ca
en ae eee)
a Or “
Pa Oe ae .
ea a
‘te * as °
es co
5 5 °
a: - oe 5 a y =
Sy
. ay . 5 7 .
. a 7 :
: oY
. 5 » . . °
a p * . ,
e a a ee s . ° ° °
. i
. j Oo
5 : a ee . ao . A
F . - « .
, ar -
: o » Co ” Oy . A .
, an A ao
5 A ns
te - 4
a . E acer 7 . bs . co
Per a oO og ,
.
. .
E : a
a 7, . .
Ps 7 o cd
ao ee s : . x
cry an S fee
4 ' ae
OCR text:
a UGS
Sha” : FA. \
: a an aa SENAY
: a)
yin
mY
; a,
OCR text:
et i
, . My personal notes!
) a" a ‘an
ia
j=
OCR text:
Hi, I'm Paul lea <_
Sie LLM Engineer's
co coal 25 a ne?
oo eS Handbook
th a wees Founder & CEO @ Decoding Al t
a Content and courses on shipping Al products :
i a cates
/ Co-Author of the ” on
LLM Engineer’s Handbook Bestseller
Pont taetin | Mewtme Letenne <packt>
¥ a. i
OCR text:
\
49 >
he
le
Vs
re,
OCR text:
; 7 ao
: =v = Hi, I'm Louis-Francois
oe = | — Co-founder & CTO @ Towards Al
PF
n — “What's Al’ on YouTube - author of Building LLMs for Production
Va . a — Previously PhD @ Mila
— Build courses, videos, and trainings for a living
TOWARDS A
o Which all starts from good research!
OCR text:
‘a [
{ aus
ad |
TOWARDS AN
, NN 4
OCR text:
(problems) Every research session starts from zero
Right now: give Codex 5 links, 3 repos, PDFs, two videos, my notes — it answers
WH PaO
w7
ie
OCR text:
| eee
oy “eg ,
i \ ; i
: e,
f
TOWARDS Ni
nn a so
OCR text:
4° agree My OS didn't start from zero
. j co a bm
: ‘~ 4
i : ye — Years of notes — every video and course I've made, all moved into Obsidian
- ; | — Meeting recaps | save after every call -> Obsidian
i,
al — Highlights from posts, tweets, and articles -> Obsidian
1" eka ; — Even my agent skills -> Obsidian
TOWARDS NA
OCR text:
One long-term research companion
for everything | teach.
and it compounds — every note, meeting, source. and question feeds it Ww BK
ww
a " &
OCR text:
v1: a topic in, research.md out
~ ON %
Ss Lf NY
i aa i
/ p Topic
OCR text:
GHA
ee Ba! iZ
Se oe | tw
in by A eee NN
rat cag . a NN ho
cial
lt £
‘ ‘ : in
: ; F
OCR text:
e : _ Forget the infrastructure you think
_ eae you need
as
——~ . aw No vector database. No knowledge graph.
\\" No semantic search. No text search.
OCR text:
No database, just the index
; ec eS.
ae mss 2=={ |
a yl catalog « summanes:
ad ae)
x ; Agent
Vr...
! “ i
OCR text:
No database, just the index
Lg
ig -\.
rN
i al a By “ Ni @ roads first index.yaml aa manees
. im a b yy catalog + summaries ee
H ayer 4z] wiki derivatives
the index is the retrieval layer N
=| raw source
~awW ga
no vector OB
OCR text:
Howdoesthewikilooklike?
research-coding-agent-architectures
raw
ARCHITECTURIARCHITECTURE
wiki
comparisons
concepts
entities
notes
repos
sources
open-questions
nstruction-ties
overview
ARCHITECTURE
synthesis
index
log
OCR text:
One immutable PARA snapshot
‘ 6K i AS. Projects
‘ / a Si Obsidian snapshot
i “ ra oe Areas immutable
| ,
|
OCR text:
©} re a rosearch-0%- workshop
(eld peers er anes oer Prey Eanes wens Seen Eee aoarares peers
at Pasty
RST e|
tebdaig ot
Pee
ct .7 2 i ny
en
a awh
a a NF
5
rs
ar
OCR text:
« ai-research-os-workshop a eS as
man f 0 Cees Coen becca
Pe ras re rae rl] ery
eee
ora
Do pap aR eke oa
Dra
ee een Tere)
CCS Ran
& abNk
Ter eames
eee
Lael eye hy
Limes Ls
of Contributors 4
cs a a Serr oo)
ere ai-research-os-workshop workshop s
a>
ii) ae ‘ ets
iL we? < booms Desa TCO HOI EV ae Tat Soe a onde ea eee ee Oe CTL oe ES - eunarnat
. 5 Cota ntes ery ue erat a y
NA meer CLt ee Cre test TED
Languages
Fors rp 2 qvesters, you sncu'd If you "ave one repo, one ca or ora Quica Question, open Codex or C-aude Cade
SIE EL Cen Oe Eee
OCR text:
Second-Bain
erompcnd
Profects/Building Your Oen AI Research OS（Empty)/examgle_1_deep_research/proe
19.Coselalon-TFtteefms
1.Z
Osls
uplicetions,eillarthildinghares.
Dals
(1stinpjoo
d fce.ealie t girg tpt te sat t f it,cpled te
giligt rt fi peiae tg ts,p
D
verycatoe ut/oandas lsgit.
3.15w
D1-tamg
esnrtes
2-Bler
D3-Meda
1.yAAton r(://etllh.c/etig/ea-ogte-y)
2.(ffectie hrnfr long-rtng eta(ht://thrpe.c/egir
D4-Joumal
/etfectise-basem-fer-log-rsag-ageta)
S-Soures
3.(lding A Age riLangChrisa e(tt:/yt./tch
D6-Nos
4isagtyMel)
D7-A
Crt t r dectryts ietery is.
8-Poec
DAlEngeersHacbook
DBdngaCodngAgntFrom Seh
DudngYsur OnAL es
rhO5(
promp.md
Dep_2.gb
Deample3ges_inks
DCortent
DHo
D.9-Ac
anv.sh
Nen
AGENTSNd
CUAUDEmS
MP
ectio
wo
OCR text:
Lc Reswneuh Septet }
Ee
¢odt testa ial reg Serene Mr ie eC
brhiasibghtoepheliniaboy one toup nowretaainid pemhmnsbic x Sam
Epp alr a SRSA PO OP
Mepis
. ad
a a
Sane a 4
nee 2A
oe ey re,
OCR text:
Newtab
Newtab
1-Templates
2-Buffer
3-Media
4-Jounal
5-Sources
6-Notes
7-Areas
8-Projects
Agentic Al Engineering Course
X
Buildinga CodingAgentFrom Scratch
Building Your OwnAl Research OS
example_1_deep_research
Create newnote （N）
research-agentic-harness
Gotofile（O）
raw
Close
wid
index
log
prompt
example_3_ingest_links
Building Your Own Al Research OS (Emp-.
Content
Moving
9-Archive
AGENTS
Second-Brain
Slide-Derived Subjects To Review
Subject extraction uses video title, related session titles/descriptions, transcript context, and OCR text when available. OCR is best-effort and should be reviewed against the embedded slide images.