News for hackers. Stuff that matters.
Re:X, the moonshot factory by prepend on Monday July 24, 2023 @12:33PM (#36850622)
How long do these “moonshots” have until they are considered failures?

It seems like an anti pattern that some huge, futzy organization launches a “moonshot” to explain why they burn money on something without needing a return because it’s a “moonshot.”

JFK gave his “to the moon” speech in 1962, targeting 1970 [0]. And the USG did it. So 8 years was the first one. It cost a ton of money and was successful and had massive returns culturally, scientifically, and financially.

Google launched its X/moonshot org in 2010 [1]. 14 years later, have any “landed?”

It seems like orgs should stop using this terminology unless they are seriously committed.

The USG launched their “cancer moonshot” [2] for very good reasons, and very sympathetic, but their goal was by 2020 and they didn’t make it. I think we should increase cancer research funding, but draping it in hyperbole is negative when we don’t achieve what we said we targeted.

[0] https://en.wikipedia.org/wiki/We_choose_to_go_to_the_Moon

[1] https://en.wikipedia.org/wiki/X_(company)

[2] https://en.wikipedia.org/wiki/21st_Century_Cures_Act#Medical...

Re:X, the moonshot factory by brycewray on Monday July 24, 2023 @12:44PM (#36850793)
The main JFK speech about the manned moon landing was May 25, 1961, before a joint session of Congress. The “we choose to go to the moon” speech at Houston’s Rice University in the fall of 1962 was more a case of pumping up public enthusiasm about the very expensive program, given that polls — especially in that mid-term election year — were showing growing doubts among voters that the expenditure was worth it.
Re:X, the moonshot factory by prepend on Monday July 24, 2023 @02:15PM (#36852117)
Thanks for the clarification. So it’s more like 9 years than 8.
Re:X, the moonshot factory by pmcjones on Monday July 24, 2023 @03:38PM (#36853193)
Apollo 11 landed on July 20, 1969, so just over 8 years after the main speech at the joint session.
Re:X, the moonshot factory by ke88y on Monday July 24, 2023 @12:49PM (#36850885)
> Google launched its X/moonshot org in 2010 [1]. 14 years later, have any “landed?”

Waymo, Brain, and Chronicle

Re:X, the moonshot factory by giobox on Monday July 24, 2023 @01:20PM (#36851312)
I can't speak to the other two, but it's laughable to suggest in 2023 Waymo has "landed" - where is the contribution to Google's bottom line? Where can I buy or use Waymo tech outside of some very restricted demo neighborhoods? What is the likelihood this changes in the next few years?

I'd be far more prepared to argue Waymo is presently another poster child for a failing Google x project.

Re:X, the moonshot factory by ke88y on Monday July 24, 2023 @01:31PM (#36851474)
> outside of some very restricted demo neighborhoods

Nearly 200 square miles of the greater Phoenix metro. If Waymo services your home address -- and it services a lot of residential area -- then you can probably get anywhere you'd want to go around Phoenix in a driverless taxi.

> What is the likelihood this changes in the next few years?

I have no insider info, but I'd expect SF and LA County to roll out in the next few years and significant coverage of major metros by the end of next decade.

Sorry driverless cars aren't happening fast enough for you, but they are happening.

Re:X, the moonshot factory by giobox on Monday July 24, 2023 @01:37PM (#36851558)
> Sorry driverless cars aren't happening fast enough for you, but they are happening.

Don't know why you are saying sorry to me either. They are happening of course, but to say its "landed" at Waymo is hilarious to me.

Re:X, the moonshot factory by ke88y on Monday July 24, 2023 @01:51PM (#36851752)
> Precisely - highly restricted demo neighborhoods. This USA is a massive country.

1. IDK why anyone would spend the money to do driverless everywhere in the USA. You can get to 30% of the population and by far the most profitable taxi markets by going one metro at a time. Even just the major metros in the sunbelt would print insane amounts of money.

2. I'm going to go out on a limb here and say that the time to go from one metro to N metros in not N*(lifetime of waymo). Ie, the hard part is done. The product has made it to market, at least in the sunbelt, and the poo-poo attitude about being able to go anywhere in the Phoenix metro without a driver is a pretty amazing dismissal.

Re:X, the moonshot factory by giobox on Monday July 24, 2023 @01:53PM (#36851778)
> You can get to 30% of the population and by far the most profitable taxi markets by going one metro at a time. Even just the major metros in the sunbelt would print insane amounts of money.

When this actually happens, then we can discuss Waymo's "landing". This is all hypothetical today.

It's not about traversing the whole USA - its about serving more than a handful of customers in demo areas, relatively speaking. Waymo has never done this to date.

I'm not disputing Waymo could go somewhere - it might! It hasn't so far.

Re:X, the moonshot factory by ManuPop on Tuesday July 25, 2023 @07:28AM (#36860746)
Describing “landed” as “Waymo is profitable and deployed in most of the US” would be like arguing that the space program was a failure because it didn’t lead to profitable space ports and space routes.

The equivalent of the moon landing for self driving tech is closer to “they have driven a lot of miles without anybody in the driver seat”, which both Cruise and Waymo have done (>1M miles)

Re:X, the moonshot factory by giobox on Tuesday July 25, 2023 @11:53AM (#36864195)
What utter nonsense. Google is a publicly traded for-profit company, not a sovereign government or charitable cause.

The entire point of Google's moonshots is new profitable businesses, of which Waymo is not and has never been. Comparing to the United States Government because they both used the term "moonshot" is beyond silly.

For a Google moonshot to land, it has to make money, or it won't continue to exist. It's really not a complex concept.

Re:X, the moonshot factory by sf_rob on Monday July 24, 2023 @01:30PM (#36851460)
Even though it came out of X, I don't see how Chronicle could reasonably be considered a moonshot for Google. A cloud-based SecOps suite that emphasizes integration with Google Cloud/GSuite seems like a fairly obvious vertical expansion. That said, I don't know much about SecOps.
Re:X, the moonshot factory by ke88y on Monday July 24, 2023 @01:34PM (#36851514)
I'd argue that the OG Moonshot was less of a metaphorical moonshot than most of the X projects. Self-driving is obviously WAY harder than going to the moon, and I think we were way further away from a science and tech perspective than we were from the moon in the early 60s.
Re:X, the moonshot factory by prepend on Monday July 24, 2023 @02:14PM (#36852108)
> Self-driving is obviously WAY harder than going to the moon

I don’t think that’s fair in the sense that going to the moon was absurdly difficult in 1962. It’s hard to compare things so different, but going to the moon only seems easier now, 60 years later. At the time of the moonshot it was one of the most difficult things humanity ever attempted.

Re:X, the moonshot factory by julianlam on Monday July 24, 2023 @01:01PM (#36851065)
I haven't heard of the latter two, and Waymo is currently in development/testing and is not nearly as ubiquitous as it seems to be in the States.
Re:X, the moonshot factory by tehlike on Monday July 24, 2023 @01:02PM (#36851086)
Brain is everything you know of today in GenAI space, and many more internally used products and technologies.
Re:X, the moonshot factory by sparrc on Monday July 24, 2023 @01:09PM (#36851161)
How much of that was actually generated directly from "Google X"? Seems to me that most of it came out of DeepMind, which was just a run-of-the-mill big company buys innovative startup type of acquisition.
Re:X, the moonshot factory by ke88y on Monday July 24, 2023 @01:17PM (#36851279)
Brain and DeepMind were separate orgs until very very recently. Google's GenAI stuff was mostly out of Brain and its RL stuff was mostly out of DeepMind.
Re:X, the moonshot factory by foooorsyth on Monday July 24, 2023 @01:16PM (#36851267)
>Brain is everything you know of today in GenAI space

I know ChatGPT, Midjourney, StableDiffusion, Adobe Firefly, And GitHub Co-Pilot in this space. Where’s Google?

Let me guess: Brain’s research led to those products? Even if that were the case, Google still failed to market and capitalize. Are they running an open research lab for the greater good? Or are they still interested in launching market-leading products?

Re:X, the moonshot factory by ke88y on Monday July 24, 2023 @01:56PM (#36851825)
Unlike the original moonshot, which was famously first-order revenue-positive ;)
Re:X, the moonshot factory by prepend on Monday July 24, 2023 @02:13PM (#36852078)
Yes, it was. It was a political program to gain reputation against the Soviets and phenomenally successful in that I think the US was happy with its ROI.
Re:X, the moonshot factory by ke88y on Monday July 24, 2023 @02:19PM (#36852175)
Reputation isn't revenue.

You could say the same thing about Brain or Waymo.

Re:X, the moonshot factory by bostonpete on Monday July 24, 2023 @03:50PM (#36853318)
The USG isn't setting out to make money on major ventures. Publicly traded companies are.
Re:X, the moonshot factory by tehlike on Wednesday July 26, 2023 @12:31AM (#36873113)
Sure, google didn't capitalize on their tech before others did.

But google is a fast follower. Let's wait and see.

Also - did you miss "many other internally used tools/infra" part - if they can feed ads infra and increase revenue by a meager 1%, that's more than enough to make up for all others combined in revenue generated.

Re:X, the moonshot factory by prepend on Monday July 24, 2023 @02:08PM (#36852007)
Has google landed anything from AI yet?

Bard just launched and it may be a $100B product in a few years, but now it’s still playing catch-up to OpenAI (that started without any moonshot and “only” with $1B in its first major seed round [0]).

[0] https://www.crunchbase.com/organization/openai

Re:X, the moonshot factory by valval on Monday July 24, 2023 @05:38PM (#36854571)
All SOTA generative models are built on Google Brain’s research.
Re:X, the moonshot factory by rand1239 on Monday July 24, 2023 @12:56PM (#36850986)
Do any of them generate profit?
Re:X, the moonshot factory by tehlike on Monday July 24, 2023 @01:03PM (#36851090)
Brain is indirectly profitable in terms of AI backend for the whole company, which pays dividends in ads, mobile, etc.
Re:X, the moonshot factory by ke88y on Monday July 24, 2023 @02:06PM (#36851975)
Did the moon landing?
Re:X, the moonshot factory by beepbopboopp on Monday July 24, 2023 @02:08PM (#36852014)
Waymo was acquired externally.
Re:X, the moonshot factory by damontal on Monday July 24, 2023 @12:54PM (#36850941)
I thought moon shot was in reference to the rescue effort for Apollo 13.
Re:X, the moonshot factory by philwelch on Monday July 24, 2023 @01:47PM (#36851689)
I think in retrospect, we can consider the Manhattan Project another moonshot, at least purely on a technical level, and that only took three years. The thing about moonshots is that you can't just declare a moonshot by fiat and expect anything to happen. The state of science and technology needs to have reached a point where a significant breakthrough is imminent, given sufficient investment. This is where atomic science was in 1942 and rocketry was in 1960.

Which is why the historic examples of moonshots didn't actually arise from government fiat, but from scientists. The Manhattan Project famously started from a letter, signed by multiple prominent physicists including Einstein, warning the US government that the invention of an atomic bomb was imminent enough that they feared Germany would accomplish it first. The moon landing was a little more complicated, but for the entire 20th century, rocket scientists understood it was possible to explore space and did everything they could to try and convince various governments to fund their research[1]; the imminent achievability of a moon landing by 1960 fueled fears in the US government that the Soviets might beat them there.

So who are the cancer researchers promising that any sort of "moonshot" breakthrough is imminent? Let me know if there is one, but the cancer moonshot idea seems to have originated in politics, and once politicians start promising "cancer moonshots"--even if no such thing is likely to be achievable--it's not likely cancer researchers are going to complain or refuse the new source of potential funding.

Unlike the government, Google at least has or had the potential to be the type of organization that actually listened to the relevant technical experts instead of being dictated by the whims of random executives receiving revelations from the Good Idea Fairy. But I agree with you that it doesn't seem like they've delivered.

Finally, I think even calling these projects "moonshots" sets them up for failure in a certain way. If you said "we can and will land on the Moon" in 1962, people will be impressed. If you say "we're going to achieve a 'cancer moonshot'", what does that mean? Does that mean you're going to cure cancer? No, because if you were going to cure cancer, you'd say "we can and will cure cancer". Except (as far as I know) that would be a lie, so instead they say "cancer moonshot", which doesn't mean anything. Any goal that's ambitious enough to be called a "moonshot" is a goal that you can just say out loud and it will impress people a lot more than calling it a "moonshot".

[1] Yes, this is my way of glossing over the V-2; I don't morally endorse Von Braun's work during the war and this is entirely beside the point I'm making here.