Peter Som de Cerff
@psomdecerff.bsky.social
@psomdecerff.bsky.social He/Him
Red-state urbanist, sometimes cyclist, walkable city proponent.
Systems engineer, safe systems advocate, rail supporter.
Working collectively to make cities better for people: more housing, more transit, fewer cars!
Red-state urbanist, sometimes cyclist, walkable city proponent.
Systems engineer, safe systems advocate, rail supporter.
Working collectively to make cities better for people: more housing, more transit, fewer cars!
This is at the root of much of the problem, but it's not so much causal as correlated, as part of evolving feedback loops. Western judeo-christians are relative wealthy, and Muslims tend to be poor despite vast resources, esp since extractive-industry wealth tends to promote heavy inequity.
June 13, 2025 at 3:21 PM
This is at the root of much of the problem, but it's not so much causal as correlated, as part of evolving feedback loops. Western judeo-christians are relative wealthy, and Muslims tend to be poor despite vast resources, esp since extractive-industry wealth tends to promote heavy inequity.
#ArtificialIntelligence
Given that AIs are rapidly improving on their performance on HLE (new a month ago, about?), does this imply that the next AI benchmark will need to developed by an AI?
Given that AIs are rapidly improving on their performance on HLE (new a month ago, about?), does this imply that the next AI benchmark will need to developed by an AI?
February 15, 2025 at 10:11 AM
#ArtificialIntelligence
Given that AIs are rapidly improving on their performance on HLE (new a month ago, about?), does this imply that the next AI benchmark will need to developed by an AI?
Given that AIs are rapidly improving on their performance on HLE (new a month ago, about?), does this imply that the next AI benchmark will need to developed by an AI?
#ArtificialIntelligence
Systems Engineering Minute: Narrow AI?
I've been traveling a good bit, and haven't much posted. While working an AI slide for work, I happened across a nice graph, and I though it worth passing along.
Systems Engineering Minute: Narrow AI?
I've been traveling a good bit, and haven't much posted. While working an AI slide for work, I happened across a nice graph, and I though it worth passing along.
February 15, 2025 at 9:57 AM
#ArtificialIntelligence
Systems Engineering Minute: Narrow AI?
I've been traveling a good bit, and haven't much posted. While working an AI slide for work, I happened across a nice graph, and I though it worth passing along.
Systems Engineering Minute: Narrow AI?
I've been traveling a good bit, and haven't much posted. While working an AI slide for work, I happened across a nice graph, and I though it worth passing along.
#urbanism #congestionpricing
Pretty sure this is the busiest Holland Tunnel has been this year.
Guess what? It's not commuters...but people going to the city for Friday night.
Pretty sure this is the busiest Holland Tunnel has been this year.
Guess what? It's not commuters...but people going to the city for Friday night.
January 18, 2025 at 1:53 AM
#urbanism #congestionpricing
Pretty sure this is the busiest Holland Tunnel has been this year.
Guess what? It's not commuters...but people going to the city for Friday night.
Pretty sure this is the busiest Holland Tunnel has been this year.
Guess what? It's not commuters...but people going to the city for Friday night.
Check it out - new latest-day overlay on top of the Pre-CP and Post-CP averages!
www.congestion-pricing-tracker.com
www.congestion-pricing-tracker.com
January 13, 2025 at 4:48 PM
Check it out - new latest-day overlay on top of the Pre-CP and Post-CP averages!
www.congestion-pricing-tracker.com
www.congestion-pricing-tracker.com
#urbanism #congestionpricing
The tracker now has a full week of data, and a tidbit of new analysis:
Really no surprises in the fundamentals of the observed shifts, as this is what any economics model would suggest. What is more profound is the DEGREE of observed changes.
The tracker now has a full week of data, and a tidbit of new analysis:
Really no surprises in the fundamentals of the observed shifts, as this is what any economics model would suggest. What is more profound is the DEGREE of observed changes.
January 12, 2025 at 3:38 PM
#urbanism #congestionpricing
The tracker now has a full week of data, and a tidbit of new analysis:
Really no surprises in the fundamentals of the observed shifts, as this is what any economics model would suggest. What is more profound is the DEGREE of observed changes.
The tracker now has a full week of data, and a tidbit of new analysis:
Really no surprises in the fundamentals of the observed shifts, as this is what any economics model would suggest. What is more profound is the DEGREE of observed changes.
Yup, $9 CP was clearly effective today. Won't really know HOW effective without actual traffic counts.
January 12, 2025 at 1:40 AM
Yup, $9 CP was clearly effective today. Won't really know HOW effective without actual traffic counts.
#urbanism #congestionpricing
Systems Engineering Minute - Another Queuing Example
Let's look at another traffic graph showing congestion. This is a typical Saturday for the Holland Tunnel, pre-CP. Consider our toolset from yesterday:
- Nominal delay is 8 minutes - the time it takes to traverse
Systems Engineering Minute - Another Queuing Example
Let's look at another traffic graph showing congestion. This is a typical Saturday for the Holland Tunnel, pre-CP. Consider our toolset from yesterday:
- Nominal delay is 8 minutes - the time it takes to traverse
January 11, 2025 at 2:44 PM
#urbanism #congestionpricing
Systems Engineering Minute - Another Queuing Example
Let's look at another traffic graph showing congestion. This is a typical Saturday for the Holland Tunnel, pre-CP. Consider our toolset from yesterday:
- Nominal delay is 8 minutes - the time it takes to traverse
Systems Engineering Minute - Another Queuing Example
Let's look at another traffic graph showing congestion. This is a typical Saturday for the Holland Tunnel, pre-CP. Consider our toolset from yesterday:
- Nominal delay is 8 minutes - the time it takes to traverse
Often it is probably over-capacity, with drivers pushing limits, which makes it ripe for congestion collapse and crashes.
So let's look again at today's congestion spike:
So let's look again at today's congestion spike:
January 11, 2025 at 2:15 AM
Often it is probably over-capacity, with drivers pushing limits, which makes it ripe for congestion collapse and crashes.
So let's look again at today's congestion spike:
So let's look again at today's congestion spike:
P.P.P.S Looking closer, and assuming a LOT, we might infer a few things from today's Holland Tunnel graph:
Something happened at about 10am - somebody corked the drain. But look at the slopes -- the rise is about half as fast as the drop after.
Something happened at about 10am - somebody corked the drain. But look at the slopes -- the rise is about half as fast as the drop after.
January 11, 2025 at 1:08 AM
P.P.P.S Looking closer, and assuming a LOT, we might infer a few things from today's Holland Tunnel graph:
Something happened at about 10am - somebody corked the drain. But look at the slopes -- the rise is about half as fast as the drop after.
Something happened at about 10am - somebody corked the drain. But look at the slopes -- the rise is about half as fast as the drop after.
P.P.S. Here on the Lincoln Tunnel, we see daily repeats of a similar, broader peak, and this is more significant. Such a pattern says that as rush hour progresses, traffic goes from ingress < egress, to ingress > egress, as the "faucet turns up", but the relatively short peak and rapid decay...
January 10, 2025 at 10:39 PM
P.P.S. Here on the Lincoln Tunnel, we see daily repeats of a similar, broader peak, and this is more significant. Such a pattern says that as rush hour progresses, traffic goes from ingress < egress, to ingress > egress, as the "faucet turns up", but the relatively short peak and rapid decay...
P.S. Take a look at this tunnel graph - and the short spike of congestion. Either a LOT of added cars ingressed the tunnel, or somebody "stoppered the drain" with a wreck.
That the congestion dropped quickly once the stopper opened implies that egress >>> ingress,...
That the congestion dropped quickly once the stopper opened implies that egress >>> ingress,...
January 10, 2025 at 10:31 PM
P.S. Take a look at this tunnel graph - and the short spike of congestion. Either a LOT of added cars ingressed the tunnel, or somebody "stoppered the drain" with a wreck.
That the congestion dropped quickly once the stopper opened implies that egress >>> ingress,...
That the congestion dropped quickly once the stopper opened implies that egress >>> ingress,...
...are flowing better. And not just a little - the visible differences are stark. As an example, here is a graph of the Holland Tunnel, before CP and after, and it's pretty easy to say "wow, that's a huge difference! 80% of the cars went away!" Well, maybe....and maybe not.
January 10, 2025 at 10:18 PM
...are flowing better. And not just a little - the visible differences are stark. As an example, here is a graph of the Holland Tunnel, before CP and after, and it's pretty easy to say "wow, that's a huge difference! 80% of the cars went away!" Well, maybe....and maybe not.
#urbanism
#congestionpricing
Seeing more red this morning across the region, but the Manhattan area still has a lot of green.
Lincoln Tunnel congested more today, but Holland is still pretty good - the highways are backing up some, tho.
Still working, but you can see variable pricing could help.
#congestionpricing
Seeing more red this morning across the region, but the Manhattan area still has a lot of green.
Lincoln Tunnel congested more today, but Holland is still pretty good - the highways are backing up some, tho.
Still working, but you can see variable pricing could help.
January 8, 2025 at 12:41 PM
#urbanism
#congestionpricing
Seeing more red this morning across the region, but the Manhattan area still has a lot of green.
Lincoln Tunnel congested more today, but Holland is still pretty good - the highways are backing up some, tho.
Still working, but you can see variable pricing could help.
#congestionpricing
Seeing more red this morning across the region, but the Manhattan area still has a lot of green.
Lincoln Tunnel congested more today, but Holland is still pretty good - the highways are backing up some, tho.
Still working, but you can see variable pricing could help.
#urbanism
: "Welp, that's your problem, right there."
Almost all the surface spots are apparently still free...even in Manhattan? Is this actually true still?
If the Manhattan spots were paid, that'd be what, another $100M per month of revenue?
Almost all the surface spots are apparently still free...even in Manhattan? Is this actually true still?
If the Manhattan spots were paid, that'd be what, another $100M per month of revenue?
January 7, 2025 at 7:28 PM
#urbanism
: "Welp, that's your problem, right there."
Almost all the surface spots are apparently still free...even in Manhattan? Is this actually true still?
If the Manhattan spots were paid, that'd be what, another $100M per month of revenue?
Almost all the surface spots are apparently still free...even in Manhattan? Is this actually true still?
If the Manhattan spots were paid, that'd be what, another $100M per month of revenue?
Congestion is slightly higher than yesterday, but still not bad in the tunnels.
The regional traffic view is a bit worse than yesterday, but Manhattan seems to be in good shape.
And, my personal "CP win" - Google is not suggesting rat-runs through JC to save a few driver minutes.
The regional traffic view is a bit worse than yesterday, but Manhattan seems to be in good shape.
And, my personal "CP win" - Google is not suggesting rat-runs through JC to save a few driver minutes.
January 7, 2025 at 1:28 PM
Congestion is slightly higher than yesterday, but still not bad in the tunnels.
The regional traffic view is a bit worse than yesterday, but Manhattan seems to be in good shape.
And, my personal "CP win" - Google is not suggesting rat-runs through JC to save a few driver minutes.
The regional traffic view is a bit worse than yesterday, but Manhattan seems to be in good shape.
And, my personal "CP win" - Google is not suggesting rat-runs through JC to save a few driver minutes.
How it started:
How it's going:
Yes, it's too soon to claim victory yet, as more time is need to have significant data. No, it's def not a dead on arrival.
How it's going:
Yes, it's too soon to claim victory yet, as more time is need to have significant data. No, it's def not a dead on arrival.
January 6, 2025 at 10:03 PM
How it started:
How it's going:
Yes, it's too soon to claim victory yet, as more time is need to have significant data. No, it's def not a dead on arrival.
How it's going:
Yes, it's too soon to claim victory yet, as more time is need to have significant data. No, it's def not a dead on arrival.
#urbanism #congestionpricing
So far, so good.
No rat-runs through Jersey City suggested by Google, as the highway is faster (!!). Tunnel congestion remains muted, but of course still sub-optimal.
Would love to see bus prioritization and ingress metering still, but CP appears to be working.
So far, so good.
No rat-runs through Jersey City suggested by Google, as the highway is faster (!!). Tunnel congestion remains muted, but of course still sub-optimal.
Would love to see bus prioritization and ingress metering still, but CP appears to be working.
January 6, 2025 at 12:49 PM
#urbanism #congestionpricing
So far, so good.
No rat-runs through Jersey City suggested by Google, as the highway is faster (!!). Tunnel congestion remains muted, but of course still sub-optimal.
Would love to see bus prioritization and ingress metering still, but CP appears to be working.
So far, so good.
No rat-runs through Jersey City suggested by Google, as the highway is faster (!!). Tunnel congestion remains muted, but of course still sub-optimal.
Would love to see bus prioritization and ingress metering still, but CP appears to be working.
So far it's working well, at least for reducing tunnel congestion. For anybody making >$50/hr, the 10m saved pays the toll, and for others they probably are money ahead on PATH.
The real winners are likely bus riders stuck in tunnel car traffic.
The real winners are likely bus riders stuck in tunnel car traffic.
January 6, 2025 at 12:15 PM
So far it's working well, at least for reducing tunnel congestion. For anybody making >$50/hr, the 10m saved pays the toll, and for others they probably are money ahead on PATH.
The real winners are likely bus riders stuck in tunnel car traffic.
The real winners are likely bus riders stuck in tunnel car traffic.
Why rely on anecdata when there is also data?
January 5, 2025 at 11:00 PM
Why rely on anecdata when there is also data?
I just have Google - tons of articles about this. It's a bit confusing to deconflict commuters who live in Mnh and commute by car, vs those living elsewhere to commute TO Mnh by car. Both are low percentages, though.
wagner.nyu.edu/files/facult...
wagner.nyu.edu/files/facult...
January 5, 2025 at 5:26 PM
I just have Google - tons of articles about this. It's a bit confusing to deconflict commuters who live in Mnh and commute by car, vs those living elsewhere to commute TO Mnh by car. Both are low percentages, though.
wagner.nyu.edu/files/facult...
wagner.nyu.edu/files/facult...
In coming days I'll continue to look for rat-runs suggested by Google through JC to access the Holland Tunnel a few minutes faster.
Today, it's about as good as it gets, with just a big of tunnel congestion as the highways dump onto the wide surface stroads leading to the toll plaza and tunnel.
Today, it's about as good as it gets, with just a big of tunnel congestion as the highways dump onto the wide surface stroads leading to the toll plaza and tunnel.
January 5, 2025 at 4:26 PM
In coming days I'll continue to look for rat-runs suggested by Google through JC to access the Holland Tunnel a few minutes faster.
Today, it's about as good as it gets, with just a big of tunnel congestion as the highways dump onto the wide surface stroads leading to the toll plaza and tunnel.
Today, it's about as good as it gets, with just a big of tunnel congestion as the highways dump onto the wide surface stroads leading to the toll plaza and tunnel.
Thank you for PATH improvement advocacy!
January 5, 2025 at 4:14 PM
Thank you for PATH improvement advocacy!
But why does this simple section of fairly wide (6 lane?) road become a weekend choke-point?
January 4, 2025 at 6:43 PM
But why does this simple section of fairly wide (6 lane?) road become a weekend choke-point?
#jerseycity
Anybody know why there seems to often be congestion on I78 and HWY 139 on the run to the Holland Tunnel on weekends?
Often at rush-hour the tunnel congests from the far-side (a policy and design problem), and sometimes maybe due to traffic in the tunnel (a traffic-load-mgmt issue).
Anybody know why there seems to often be congestion on I78 and HWY 139 on the run to the Holland Tunnel on weekends?
Often at rush-hour the tunnel congests from the far-side (a policy and design problem), and sometimes maybe due to traffic in the tunnel (a traffic-load-mgmt issue).
January 4, 2025 at 6:43 PM
#jerseycity
Anybody know why there seems to often be congestion on I78 and HWY 139 on the run to the Holland Tunnel on weekends?
Often at rush-hour the tunnel congests from the far-side (a policy and design problem), and sometimes maybe due to traffic in the tunnel (a traffic-load-mgmt issue).
Anybody know why there seems to often be congestion on I78 and HWY 139 on the run to the Holland Tunnel on weekends?
Often at rush-hour the tunnel congests from the far-side (a policy and design problem), and sometimes maybe due to traffic in the tunnel (a traffic-load-mgmt issue).