WEBVTT

00:00:00.833 --> 00:00:03.800
Paxton builds software
around real workflows.

00:00:03.800 --> 00:00:06.900
His independent projects span
a native Mac assistant,

00:00:06.900 --> 00:00:09.433
connected personal apps,
and a research tool

00:00:09.433 --> 00:00:11.533
for discretionary day traders.

00:00:11.533 --> 00:00:13.133
He owns the product decisions

00:00:13.133 --> 00:00:15.767
and the work of making
the result dependable.

00:00:17.067 --> 00:00:19.900
This is independent, AI-assisted work.

00:00:19.900 --> 00:00:23.533
Paxton uses AI to accelerate
implementation and review,

00:00:23.533 --> 00:00:25.967
while taking responsibility
for architecture,

00:00:25.967 --> 00:00:28.267
testing, and release decisions.

00:00:28.267 --> 00:00:29.600
The common thread is making

00:00:29.600 --> 00:00:32.967
complicated tasks
understandable and useful.

00:00:34.067 --> 00:00:37.267
Jarvis connects voice,
screen context, and memory

00:00:37.267 --> 00:00:39.433
in a local-first Mac assistant.

00:00:39.433 --> 00:00:42.167
Its native Swift interface
and Python service

00:00:42.167 --> 00:00:45.100
connect everyday requests
to specific capabilities,

00:00:45.100 --> 00:00:48.433
giving the conversation
a practical role on the desktop.

00:00:49.500 --> 00:00:51.700
The interface makes its state visible.

00:00:51.700 --> 00:00:54.500
A compact presence expands
as Jarvis wakes,

00:00:54.500 --> 00:00:58.933
then changes as it listens,
reasons, observes, and responds.

00:00:58.933 --> 00:01:01.933
This preview uses
the current native animation code,

00:01:01.933 --> 00:01:05.367
including its response aura
and return to standby.

00:01:06.467 --> 00:01:08.533
Speech recognition handles input.

00:01:08.533 --> 00:01:11.333
Authorized screen capture
adds context.

00:01:11.333 --> 00:01:13.867
Local memory preserves
useful information,

00:01:13.867 --> 00:01:16.567
and different tasks
follow different model routes.

00:01:16.567 --> 00:01:18.800
Together, those pieces
connect a request

00:01:18.800 --> 00:01:22.133
to the right information
and the next useful action.

00:01:23.167 --> 00:01:25.700
Paxton also worked
on interrupted audio,

00:01:25.700 --> 00:01:29.767
changing permissions, device handoffs,
and incomplete actions.

00:01:29.767 --> 00:01:32.167
Visible confirmation
and bounded recovery

00:01:32.167 --> 00:01:35.767
keep an uncertain result
from becoming an automatic repeat.

00:01:36.800 --> 00:01:38.900
Personal Trainer connects
food capture,

00:01:38.900 --> 00:01:41.700
nutrition, training, recovery,
and journaling

00:01:41.700 --> 00:01:43.167
in a private product.

00:01:43.167 --> 00:01:45.667
Its iPhone app
and responsive web experience

00:01:45.667 --> 00:01:48.733
organize the day
around one connected record.

00:01:49.800 --> 00:01:52.267
Food capture starts
with on-device label reading

00:01:52.267 --> 00:01:53.833
and barcode references.

00:01:53.833 --> 00:01:55.567
The person reviews the product,

00:01:55.567 --> 00:01:57.500
then chooses the amount eaten.

00:01:57.500 --> 00:02:00.467
Saving a food stays separate
from logging a meal.

00:02:00.467 --> 00:02:02.733
Missing nutrition values
remain unknown

00:02:02.733 --> 00:02:04.900
until there is evidence to fill them.

00:02:05.933 --> 00:02:08.233
Today, Log, Progress,
and Journal connect

00:02:08.233 --> 00:02:12.633
meals, workouts, hydration, sleep,
and personal observations.

00:02:12.633 --> 00:02:15.867
Paxton joins interface design
with owner-scoped records,

00:02:15.867 --> 00:02:18.600
reviewable saves,
and optional AI assistance

00:02:18.600 --> 00:02:20.267
with a specific role.

00:02:21.300 --> 00:02:24.367
Treadway is a separate
daily-planning and habit product.

00:02:24.367 --> 00:02:27.867
A day's markers, hydration, progress,
and a short brief

00:02:27.867 --> 00:02:30.433
come together
in a calm mobile experience.

00:02:30.433 --> 00:02:33.133
The interface makes
the next useful step prominent

00:02:33.133 --> 00:02:35.167
and the day easy to scan.

00:02:36.200 --> 00:02:38.467
Paxton carried
Treadway's design language

00:02:38.467 --> 00:02:41.067
into Personal Trainer
as a sibling product.

00:02:41.067 --> 00:02:44.200
They remain separate applications
with separate records.

00:02:44.200 --> 00:02:47.433
Together, they show how he develops
an interaction system

00:02:47.433 --> 00:02:50.533
and adapts useful ideas
to a different job.

00:02:51.567 --> 00:02:55.067
The AI work extends
beyond features inside an app.

00:02:55.067 --> 00:02:58.333
Paxton defines the task,
directs implementation,

00:02:58.333 --> 00:03:01.400
and uses independent review
to challenge the result.

00:03:01.400 --> 00:03:04.300
Tests and clear acceptance checks
turn a fast draft

00:03:04.300 --> 00:03:06.400
into something he can evaluate.

00:03:07.433 --> 00:03:10.033
His tool workflows
gather relevant context,

00:03:10.033 --> 00:03:12.500
build a change,
review it independently,

00:03:12.500 --> 00:03:14.300
and verify the outcome.

00:03:14.300 --> 00:03:17.500
That is the practical role
of AI orchestration here:

00:03:17.500 --> 00:03:19.300
reduce repeated work while keeping

00:03:19.300 --> 00:03:21.900
responsibility and decisions clear.

00:03:22.967 --> 00:03:24.400
Edge Lab brings that approach

00:03:24.400 --> 00:03:26.333
to discretionary day trading.

00:03:26.333 --> 00:03:28.833
It is a private Mac
research and review tool

00:03:28.833 --> 00:03:30.700
for preserving what a trader saw,

00:03:30.700 --> 00:03:33.000
what they planned,
what they chose to do,

00:03:33.000 --> 00:03:35.033
and what they can learn afterward.

00:03:36.100 --> 00:03:38.767
A chart image alone
loses meaning quickly.

00:03:38.767 --> 00:03:41.233
Edge Lab links observations to context

00:03:41.233 --> 00:03:44.133
and preserves the plan
before the result is known.

00:03:44.133 --> 00:03:45.667
Its journal includes decisions

00:03:45.667 --> 00:03:48.667
to trade, pass, cancel, or stay out,

00:03:48.667 --> 00:03:51.700
so review covers
more than completed positions.

00:03:52.767 --> 00:03:55.267
A winning outcome
can hide a weak decision.

00:03:55.267 --> 00:03:58.200
A losing outcome
does not tell the whole story.

00:03:58.200 --> 00:04:00.267
Edge Lab supports
reviewing the process

00:04:00.267 --> 00:04:01.933
before revealing the result,

00:04:01.933 --> 00:04:04.400
while preserving corrections
as added history

00:04:04.400 --> 00:04:07.000
instead of replacing
the original record.

00:04:08.100 --> 00:04:11.500
Live, replay, and historical research
stay separate.

00:04:11.500 --> 00:04:13.300
A bounded evidence packet can go

00:04:13.300 --> 00:04:16.333
to an independent AI reviewer
for criticism.

00:04:16.333 --> 00:04:17.933
The trader keeps the decision;

00:04:17.933 --> 00:04:20.800
the reasoning stays
available to inspect.

00:04:21.833 --> 00:04:23.400
Paxton sees a future product

00:04:23.400 --> 00:04:25.800
for discretionary traders
in this workflow.

00:04:25.800 --> 00:04:27.633
Today, Edge Lab demonstrates

00:04:27.633 --> 00:04:30.100
structured trading research and review.

00:04:30.100 --> 00:04:33.067
A commercial release
remains a future direction.

00:04:34.100 --> 00:04:36.900
Across these projects,
the strength is the full path:

00:04:36.900 --> 00:04:39.167
understand the problem,
shape the experience,

00:04:39.167 --> 00:04:41.733
connect the systems,
and check the details.

00:04:41.733 --> 00:04:45.367
Paxton works across
Swift, Python, and TypeScript.

00:04:46.633 --> 00:04:48.567
For a team turning
an unclear workflow

00:04:48.567 --> 00:04:52.500
into a focused product, automation,
or tested implementation,

00:04:52.500 --> 00:04:54.367
this is the work to explore.

00:04:54.367 --> 00:04:57.267
Visit the portfolio
for projects, case studies,

00:04:57.267 --> 00:04:59.400
and a conversation with Paxton.