Comment by wxw | Hacker News Reader

Comment by wxw | Hacker News Reader

Parent Full thread

> For this study, the Google Maps algorithm was modified to prefer alternative routes with similar travel times and segment types, effectively guiding trips away from the pre-selected congested segments

> Over a six month period, we adopted a city-wide switchback (also known as crossover) experimental design, alternating between this treatment and the control (unaltered) routing algorithm over consecutive days to appropriately measure the effect of this intervention

> Averaged across cities, we observe a median increase of around 2% in driving speeds on targeted segments, corresponding to a median decrease of 0.5% to 1.0% in fuel consumption rates

The cities were: Atlanta, Boston, Chicago, Los Angeles, Miami, New York, Philadelphia, Salt Lake City, San Francisco, Seattle.

The data and code is also available (https://github.com/google-research/google-research/blob/mast...) from the paper (https://www.nature.com/articles/s44284-026-00443-x).

Kudos Google! Nice to see this kind of work. That said, let's just build more trains?

I read the article and didn't see anywhere if they collaborated with city planners, or just ran this on their own. I get the techno-optimism of making traffic 'better' and also it may be concerning if trucks are being routed through neighborhoods where planners didn't build for it.

Do you think Google should get government sign-off on every change to its routing algorithm?

That's not a bad idea, since we pay for the roads, not Google.

Medical/psychological experiments involving human subjects are mandated by federal law to run through a review board. Did that happen here? This isn't a medical experiment, but I don't think its nuts to feel like maybe there should have been some oversight.

I fear you may overestimate the median benefit of government oversight.

I fear you underestimate the malice of the average tech company

This is only true for experiments funded by the government. In this case the experiment was funded solely by Google.

IRB approval is required for human subject research funded by private parties, too; not just government. But only really in the medical field.

>But only really in the medical field.

Looking it up, the exception where it is required is specifically for clinical investigations regulated by the Food and Drug Administration.

Does this apply to Netflix, Amazon, Disneys, and many other streaming providers's "What to watch next? Recommended..." of sorts (or music/podcast streaming services)...

Routing traffic affects not only the drivers using Google Maps, but also everyone around them IN REALTIME.

Netflix personalized recommendations affect only the individual, not those around them who are not using Netflix.

Oh yes, good point! Stand corrected!

You're acting in bad faith attaching "every change" to what the parent said.

But this isn't that crazy? Most businesses require government approval for changes that have community effects... that's the entire point of government. Stuff like construction permits and regulation around structures, water usage, pollution, building architecture. Mining companies are required to assess environmental impacts. Companies making RF devices need to get confirmation on how the devices affect people and other devices.

And that's not even considering market share/monopoly status, which also regularly invites more scrutiny.

When the government proscribes discretionary permitting to those items you mentioned corruption and scarcity follow. It is exorbitantly expensive and difficult to build a house, a powerplant, or a mine in this country. For which we all currently pay for in cost of living.

Does the tempo matter?

Like compare changing live routing to publishing a paper map showing alternate routes.

In either case it seems to me that the vehicle operator should be most responsible for the impact of the route that they use.

I think that's already very much a problem, and given that they constrained it to routes that didn't have a significantly different drive time the risk seems low.

I had a similar question and am generally curious if google has an internal IRB or equivalent for these studies. Given their outsized influence, there is a reasonable chance of uncoordinated interaction between their efforts and other ongoing studies. For a trivial example, consider interaction between google maps rerouting experiment near an event and localized uber surge pricing.

This is also my experience. It would suggest routes that make no sense (to a local) in an attempt to “divert” away from traffic, which ends up costing more time and petrol anyway.

City planners and zoning have been a disaster for US cities.

That's not what the parent is talking about. You're using the strawman of what appears to be occupancy zoning or something similar to argue that a techno fascists should have free reign to manipulate the mass movements of people within a city for their own purposes.

Generally you need to be pretty competent to get to the level of where these Google maps guys are.

To get to the competence level of a local traffic planner, you need to have stuck it out through a Civ E degree and maybe went through the motions at a firm for for 15 years before getting laid off and then wanting a "cruise control" job with the power you never would have achieved in private industry for the rest of your career.

You end up dealing with guys where you live or die based on the opinion of someone who hasn't bothered to learn anything new since 1991, and have no interest in breaking their "cruise control".

Still a strawmam somehow?

Found the European :)

The south SF Bay Area cities of Los Gatos and Monte Sereno have persistent problems with traffic gridlock during summer weekends and holidays due to navigation apps routing drivers off of southbound SR-17 and onto surface streets. It gets so bad that residents literally can't get out of their own driveways and emergency response vehicles can't move.

There actually did used to be train service from San Jose to Santa Cruz but it was abandoned in 1940 due to high costs and lack of use. Rebuilding it would cost >$1B.

https://www.goodtimes.sc/isnt-train-san-jose/

It can’t be helped. Roads are not only locally funded and the only thing allowing residents to free-ride on everyone else’s contribution to this commons is the information asymmetry. It’s similar to the fact that San Franciscans and Athertonians have “slow streets / no through streets” which allow a few to retain exclusive road use for themselves.

With increased information comes more efficient utilization and no doubt the free-riders will find that troubling since it’s no longer a free ride with the asymmetric effect where they can use my roads but I can’t use theirs.

In an ideal world, we would use our Fastraks to charge for marginal road use and a locality can recoup capital investments it made by charging appropriately for marginal use.

Do you not have shoulder lanes? I ask because I’ve lived places that don’t and I’ve always wondered if you just die if EMS doesn’t have a lane to get to you.

There are no shoulder lanes in residential and downtown areas.

> let's just build more trains?

"just" is doing a lot of work there. Trains are a very (very) expensive way to move people between pre-determined places. We can do a lot better than this in 2026.

Any examples?

Reply on news.ycombinator.com