81 comments

[ 3.4 ms ] story [ 147 ms ] thread
I can't help but feel that this reflects a change in the mindset of Google – getting further and further from its scrappy upstart roots which are now long-since gone...
RIP Google Sets :(
The logic behind Sets actually lives on, at least for now, within the autofill feature of Google Spreadsheets. Fill a cell or two with sample data, then option-drag (control-drag on Windows) the square autofill handle.
Yes, or if you search for "Oregano" you will notice the auto-complete for subsequent searches strongly favours herbs and spices.
For some reason the results from it just weren't that great anyway. It could never find sets of things that I really wanted to get.
My experience was that it started out great. I found great recommendations for books, bands, foods - even drugs. But then they tweaked it. They changed something in the settings, or maybe their index changed as they grew, and it simply stopped working. Somehow it managed to become too restrictive and too general at the same time. IMHO, sets was one of the greatest things google ever made, but it died a long time ago.
I used to use this for server names. Sad to see it go.
What is happening to Google? I almost feel like it's going through a transformation like Walter White.
They're getting ready to cook up some new projects ;)
Wouldn't Labs be a good place to link to them?
I get the feeling that the bean counters are taking over Google and slashing off everything that doesn't have obvious monetization potential.

I mean, I know great people who work there, but overall it seems like they're turning into Microsoft circa 2000 - they own the industry, they buy startups and kill the products made by about half of them. Some of their products are great, but a lot of it is just an obvious monopoly enabled money grab.

I wish this wasn't the case.

You would never seen something like "Labs" from Apple, and everyone applauds them for "focus".
Apple did have "labs." One of the first thing Steve Jobs did when he returned was shut it down.

http://en.wikipedia.org/wiki/Advanced_Technology_Group

That explanation misses a lot of context. It comes across as a criticism, as if Jobs just didn't see the benefit of research. That's not the case. He wanted to shrink the things they focused on, so they could put their 'A' team on every project (which required less projects). The end result was having a bunch of really talented guys focusing on things that drove business value.
I didn't intend it to be a criticism, just a factual remark. Then again, I shouldn't have really compared the two. Google labs wasn't really a dedicated research wing at Google, more like cool stuff people from various teams worked on in their spare time.
Of course, Apple shut the ATG down in the context of the company bleeding money and having maybe 90 days of life in it at then-current spending levels.

The optics in this case are somewhat different.

Anyone who has been to WWDC and seen their engineering staff present will know that they're given a bit of leeway in the design of the products they work on.

Real artists ship - the expansion and creation of new API's in their software shows this.

Just because they can't work on whatever with 20% of their time doesn't mean it's a no-creativity nose to the grindstone environment.

I'm about as much of a Google fanboy as you can be and unfortunately zdw's words ring true.

Also, in spite of Apple's success I'm not sure Google would do well to emulate them. I (and I would guess many others) have loved Google because they aren't Apple, though unfortunately they do seem to be inching that way at an increasing rate.

Are you talking about their recent UX trend of turning everything into a minimalist interface with a lot of whitespace?
No, I was talking more about their business side -- killing off everything that isn't a direct profit center and moves towards unopenness (no Honeycomb source code, etc).

In regards to UI, I wouldn't mind if they copied Apple more there but only if they copied the good (great defaults) without copying the bad (no ability for power users to modify the defaults).

Apple certainly does a lot of experimentation, but their approach differs greatly from Google's. The main reasoning behind Google Labs as far as I can determine is to expose projects to the broader world and determine which are best through some form of natural selection. A form of market research if you will.

Apple on the other hand has been very much against market research, preferring to experiment internally, killing of bad ideas before the world even hears about them and presenting a polished final product to the outside world when it's done and ready.

Both approaches certainly work, Google's works great for improving their existing suite of products whereas Apple's works better for incubating new products that consumers don't even realise they want yet.

Apple has never been against market research, its only done internally with internal employees as the volunteers under NDA so we never see those workings..
Hardware requires focus. Software requires experimentation.
Steve Jobs said it himself several times: Apple is, at it's core, a software company. It just puts it in a really pretty box.
Yeah, I always thought Labs was a part of Google's identity. What next, no doodles because it's a waste of time? :-(
(comment deleted)
Doodles won't.

Doodles have become a strong reason for Google search to be used in their homepage. ("Locking in a category of users").

Btw, Orkut may be the next.

Running Orkut and Google Plus doesn't really make sense unless Orkut still has a large South American userbase that would be hard to transfer over (not sure if that is the case still).
Doodles is a successful bit of marketing though. Everyone knows what it is and maybe even look forward to them on special occasions.

What did Labs do successfully?

I feel like they almost get bored with projects and then just let them stagnate. It seems like once they launch something, that is pretty much the final state of that project. I am not sad to see them narrowing their focus a bit if it means they'll continue to improve on and advance their current offerings instead of launching a new initiative every week.

However, the way that they can just toss projects aside like this makes me wonder once again why I should bother becoming invested in Google Plus... how long will it be before I get an email telling me Plus is being shut down and I should go download all my content?

It was the fragmentation of Google that made it so amazing. So many wonderful discoveries came from experimentation, and labs was a wonderful way to organize them. Disappointing.
Notice that the shutdown of all these projects are happening on Larry's watch. Maybe it's a sharpened focus on ideas that "matter", but Google needs to try more things not less.

From today's announcement about Q3 results, they still get a huge chunk of their income from advertising. They are dominant, but without another angle they are just asking to be disrupted.

I honestly never used it, and am completely unaffected by this.

What was the most successful product that emerged out of google labs?

Maps, reader, docs
Maps was bought (http://en.wikipedia.org/wiki/Google_Maps#History) and so was Writely, which became part of Docs (http://en.wikipedia.org/wiki/Writely). Dunno about Reader.
In both cases, they were mergers of acquisitions with existing products that had already been under development internally. The version of Maps that was bought was a desktop product, not all that different from many of the other shrinkwrapped satellite mapping programs available in the late 90s, and it was only when it was fused with the web-based Google Local ideas cooking up internally that it became the innovative product people know of.

It's very hard to identify accurate parentage for anything developed within Google, because the culture (and innovation process) is such that ideas get proposed, refined, merged with other ideas, abandoned, resurrected in other forms, implemented, iterated upon, and finally launched, often by many different groups. Even Google Search (usually identified with Larry & Sergey) isn't really their creation: large chunks were developed by Scott Hassan (eGroups founder and never actually a Google employee, though he had a big chunk of stock), Craig Silverstein, and various other Stanford students and early employees, and the ranking algorithm was completely rewritten early on by Amit Singhal.

Oftentimes the press - or just curious Internet comments - wants a single person they can associate with a product, and so they'll take a name (usually the most famous) and make something up. That's not how it really works, though. Most ideas have many parents.

Google suggest which later got integrated in main page is one, I guess.
Not sure if this includes other services like Gmail Labs, but I relied on "attachment reminder" throughout college. It has saved me multiple times, including now in my professional career.
I may be wrong, but I think it's been rolled into the main product. A few times, recently, I've been prompted as to whether I've forgotten an attachment, and I don't recall ever turning that Labs feature on.
No, Gmail Labs is separate and is not being shut down.
Does anything remotely like Bell Labs exist today? Is it simply not possible with everyone tightening their belts?

If you're unfamiliar with what came out of Bell Labs, check out this list: http://en.wikipedia.org/wiki/Bell_labs#Discoveries_and_devel...

Or is it the fate of software innovation that it is to be decentralized and happen in our spare time now?

Microsoft Research?
Bell Labs was a manufacturing line for science Nobel prizes. Google Labs was a project for showcasing Gmail bells and whistles. I guess it was kind of disingenuous for Google to call it "Labs" in the first place.
The Viewpoints Research Institute fits the bill: http://www.vpri.org/

They want to replace the entire software industry, since it's all a huge hack (especially the 100m+ SLOC codebases that are way too common). Goal: a full GUI system (from kernel to applications) in 20k SLOC. Take a look at the progress reports they've made. They claim a 200 line TCP, 800 line graphics system (apparently comparable to Cairo), and lots of other neat stuff.

I've always felt BBN [1] was pretty close, not the size of Bell labs, but tons of really awesome research. They've been purchased and sold many times (now with Raytheon) but they are the company behind the original arpanet and, despite many transitions, they've always seemed to preserve their culture (Leo Beranek, nearly 100 now, still stops by from time to time) and passion for research .

[1] http://www.bbn.com/

Woah woah woah, you seem to be confusing "small experiments Google hosts publicly at labs.google.com" with R&D. Since your wiki link seems to be about publishing academic articles, here's a list of academic article produced by places that are not Bell Labs: http://research.google.com/pubs/papers.html http://research.microsoft.com/apps/dp/pu/publications.aspx

The software industry is also producing patents at an absurb rate, which is another byproduct of research (theoretically).

Yeah, I guess it was a little off-topic. I realize they're different things, but shutting down Google Labs made me think of the state of R&D in general.
The whole domain would redirect back to google.com. Last night when I went to Googlelabs.com via the newest announcement of retirement of services from google [1] was indeed the last time I visited it.

The fact that many of the labs projects are still active made it much uneasy to find those products anymore.

---

[1]: http://googleblog.blogspot.com/2011/10/fall-sweep.html

(comment deleted)
(comment deleted)
It does align nicely with Steve Yegge's post: strive to be a platform. Let people build what they need. Don't try to be everything to everybody. Less products, more services. Maybe right, maybe wrong, but Google shows that they CAN DECIDE and I applaude them for that.
Do they really need Labs now? Google has grown to a size where they can easily acquire promising young start-ups out in the wild.

One advantage of having Labs back in the day was that projects could have the backing of Google's infrastructure. Now, with cloud services like AWS, the bar is set low enough that anyone has access to this kind of stuff. How hard is it to start a YouTube in this day and age?

(comment deleted)
I think they are entirely right in shutting down this outward facing display of unfinished ideas/projects.

I also think they could have done much better than just redirecting it to google.com. Instead of trying to emulate Apple, they should acknowledge that their culture is different and very nerdy. So I think what they should have done instead is to highlight that and use "labs" to instead show off their technical prowess, much in they same vein as what Festo has done : http://www.festo.com/cms/en_corp/9617.htm

A flapping wings robot, how cool is that?!

Google could really do something similar, and the world would love them for it. Instead they look more and more like a big corporation with all the boring "common sense" and mediocrity that comes with it.

It was an experiment that met it's natural end.
Google Labs itself was an experiment? That's very meta.
RIP. Rebirth of "Google acquisition era" (Jaiku, et al.)?
Google had record profits - why are they killing all their incubators?

If they kill "20% time" next then we know their best days are over.

None of the Google engineers I've met had 20% Time projects. They said 20% Time was not common for application teams because they had more visible release deadlines than the web teams.
While there's no longer a specific Labs section, they do still have Google Code, and my guess is that you'll still find the odd experiment popping up on there every now and then (and naturally gaining publicity through word of mouth).
There are still places within Google where people can run independent projects and experimental stuff. Innovation isn't dead. :)
There was a great Labs feature I missed after they killed it: "What's around here?" All the button did was add an asterisk to the search box, and that still works.
The most disconcerting thing is that a Google search for one of the Labs projects doesn't handle this change in any elegant way. I went looking for a description of Google Sets (http://www.google.com/search?q=google+sets), but the most relevant links either redirect or give 404. The one hand doesn't know what the other is doing.