Alphabet Announces Second Quarter 2023 Results [pdf] | Hacker News Reader

Alphabet Announces Second Quarter 2023 Results [pdf]

abc.xyz

99 pointsjustinwp144 comments

> In January 2023, we completed an assessment of the useful lives of our servers and network equipment and adjusted the estimated useful life of our servers from four years to six years and the estimated useful life of certain network equipment from five years to six years. This change in accounting estimate was effective beginning in fiscal year 2023, and the effect was a reduction in depreciation expense of $966 million and $2.0 billion and an increase in net income of $752 million and $1.5 billion

Interesting accounting magic

Interesting accounting magic indeed. It seems like tricks like this make it remarkably easy to game stock price. Are there any controls on this?

The controls are that if 4 years from now they decide that the servers need replacing after all and take a write-down then shareholders have a very solid case to sue the company.

Is it not possible to go back in 4 years and explain in 4 years that the market is more competitive than expected?

How does suing a company work? Is it functionally just like a forced dividend, or is it more complicated than that?

It's like any other class action lawsuit. Nothing to do with a dividend.

But if it's the stock holders that are suing the company, and they are receiving compensation from the lawsuit based on the size of their holdings, it's a cash payment from the company to the holders, which is exactly what a dividend is.

Not all stock holders will be part of the lawsuit, and some non stock holders (say people who have already sold) can be part of the lawsuit. There can also be people who claim second and third degree damages without actually holding any stock. And there is no rule that compensation has to be in proportion to anyone's holdings. All of this will be up to the court to decide.

The latter. It is functionally closer to a speculative investment. Litigation is expensive and notoriously uncertain.

That's very unlikely. In each case the change would be backed by an analysis which justified the change, nefarious intent would be pretty difficult to prove.

Not an accountant but I assume that standards such as GAAP and IFRS cover this?

No, because they're balanced over time. It's when you don't make the charges that this becomes stock manipulation.

Not writing down investments is how Enron committed it's mark-to-market fraud.

It's not accounting magic at all.

A $1000 server with a 4 year life span is expensed at $250 a year.

A $1000 server with a 6 year life span is expensed at $167 a year.

You convert your server lifespan from the former to the latter you get $83, or 33%, savings a year.

If you (intend to) keep your servers for 6 years and do not make this adjustment, then you're doing a false accounting, not the other way around.

Hmm, if those servers are handling less traffic then their 'useful life' will be extended.

That's like being told inflation is down because you can eat less

Is that $2.0b for network equipment by increasing it's depreciation period by 20%? Meanwhile increasing the deprecation period of servers by 50% only nets $1.0b?

I had the impression that servers were vastly more expensive than network equipment. I wonder what the breakdown is for network equipment.

They also took a large one off hit from the layoffs over the last 6 months. I think the layoff hit was less than the gain on the equipment recategorization but it does sort of even out a bit.

It is not an accounting trick, it means they'll actually hold on to those servers longer than originally planned, meaning they'll buy fewer servers to replace them over the long term. That's actual savings.

The accounting "magic" is that they can claim the savings today, but that's just normal amortization.

And this will increase their server failure rate, but probably not enough to move the needle in terms of the random failure buffers they account for.

It does mean they’ll hold onto them, it doesn’t mean they won’t buy more.

The server demand doesn't depend on how many servers you have. If your demand is D, you have X servers, and you need to decommission Y due to age, you'll have to buy D - X + Y. If Y gets smaller, you buy fewer.

You build brand new data centres with the latest technology. When a new technology rolls out, new builds always get it, and you start upgrading your newest data centre first. Sometimes you never make it to your oldest data centre.

Data centres fill up. They fill up with important "production" software and associated data housed in the same DC for locality, which never moves and mustn't be shut down. After a while the data centre is full and everyone needs to put their new software in the newly built empty data centre.

Over time a data centre gets technology upgrades later or not at all (say, a new-new networking fabric is invented and starts rolling out before the new networking fabric finishes rolling out). There's a slow decay in terms of service. New data centres have on-site offices of engineers who keep everything running at maximum efficiency, older data centres have a single security guard who receives a daily printout of which machines to press the reset button on while doing their rounds. Eventually an old data centre is turned off all at once.

If they're keeping older servers around longer, they may be delaying turning off the least valuable computers, continuing planning upgrades to 'new' even after 'new-new' is ready and rolling out, or slowing down the invention of new tech. As I understand it, "decommission X many machines due to age" isn't really the model.

To be fair 4 years for a server seems like very little.

Until a few years ago each CPU generation brought so much performance improvement that you would actually save money by replacing the server, because it would cut down your operational expenses (power, rack space, cooling, ...) per unit of compute. This is not so true anymore.

Cloud really does make most computing power fungible.

This replacement schedule is what's creating the fungibility of the cloud.

3 years used to be the industry standard. Electricity costs are way larger than the cost of replacing servers

If you run a lot of Windows servers, it’s sometimes cheaper to run old stuff longer as they are raising prices on a per core basis.

I was stuck supporting one of these environments, and made a business case to overbuy the hardware a few years ago when they did a “free” socket to core migration. They’re still running those things, as the maintenance on the servers is less than paying for more cores.

so This is how they managed to pretend clout didnt make a multi hundred $million loss this quarter for the first time ever.

Ads is still growing and very profitable, YouTube is back to growing, GCP profitable for two quarters in a row, and lower losses on other bets.

Google is still Google.

I myself contributed to those ad earnings by watching 4 ads in a 10 min video.

Ditto. But I do wonder how much ad time is just people using it to listen to music in the background and not really listening to ads (maybe this is priced in anyways and it doesn't matter, but traveling around eastern europe a lot I was surprised how many people used Youtube this way)

The advertising people know that they can slip information and impressions into your brain whether you think you're listening/reading/clicking or not.

All they really need to do is tick a few % of the marginal decision-makers who need to (for example) buy a car to say "I should consider a Ford" and they've covered the ad budget.

This is why "impression" metrics are not very useful. Perhaps 99% of impressions are just wasted so advertisers prefer to pay for clicks, or even conversions.

YTDL over my dead body.

You could find ways around that .. after all this is Hacker News.

Honest question: Why not subscribe?

IIRC content creators get more revenue per user from YT premium than ads.

I pay, and it’s worth it for me, but I hate giving money to Google, and paying nearly twice what any streaming service charges is painful.

Twice? For me YT Premium is cheaper than even the cheapest Neflix plan, same for Disney and HBO, only Apple TV is cheaper (barely).

YT premium is now $13.99.

Netflix starts at $6.99.

That's exactly double if you ignore the missing pennies.

Netflix Standard is $15.49 though, if we're comparing without ads.

The $6.99/mo Netflix plan is with ads. The comparable YouTube plan is $0/mo.

The comparble Netflix plan to YT premium is $15.49/mo.

And YouTube includes music streaming.

In the UK YT Premium is £11.99, Netflix Standard (no ads) is £10.99

For comparison my 5G phone plan is £10/month, and my home fibre connection ADSL is £22/month.

YouTube premium is therefore a tough sell - with netflix you at least get loads of stuff you wouldn't have seen otherwise, YT you get the same stuff you could see anyway, just without ads.

I personally don't see that as a good deal. If it was like £2 or £3 a month to remove the ads, then great sign me up. But it's not.

Apparently I’m paying considerably extra for the ease of managing my subscription through Apple.

Don’t worry, I offset you some by installing adblockers on many older people’s phones and computers, especially people who aren’t technical and had no idea how to do it themselves.

I’m genuinely curious to know: why do some folks browse the web without using an ad blocker? That’s not a rhetorical question. I understand grandma may not be technically savvy enough. But for anyone reading HN, why aren’t you using an ad blocker?

Maybe they watched it on a TV and do not have access to an ad blocker on the Youtube app on the TV.

SmartTube is a life-saver on TV. It replaces the default youtube app and removes all the ads/sponsored sections

There are some people who see it as an ethical position: ad supported businesses need to serve ads to survive. If you don't want to see those ads, and the business doesn't offer a subscription plan, just don't go to those websites.

> just don't go to those websites

I know this is Youtube, but in general: I would gladly avoid those sites but they are gaming the SEO. They are getting in the way.

Ad blockers don't work on phones, which is increasingly the medium of consumption.

Brave works just fine on phones...

Ad blockers currently work on Android devices[1][2][3][4], preferably by gaining root on the device and using a de-Googled Android, using F-Droid[1] and/or manually installing packages with FFUpdater[1].

If you really have to use Apple devices and other post-general-purpose-computing locked down devices (e.g. televisions), you could try self-hosting privaxy[5][6] (or similar) which performs MITM manipulation of web traffic to remove ads, including via JS or CSS injection.

[1] https://f-droid.org/en/packages/de.marmaro.krt.ffupdater/

[2] https://play.google.com/store/apps/details?id=org.mozilla.fi...

[3] https://github.com/mozilla-mobile/firefox-android

[4] https://addons.mozilla.org/en-US/android/addon/ublock-origin...

[5] https://github.com/Barre/privaxy

[6] https://news.ycombinator.com/item?id=31432848

On android, just use firefox. Pretty enjoyable if you just switch to firefox everywhere.

GCP is an interesting one -- I helped a company go through some of the hoops for GCP and AWS and the former was... difficult. It felt like someone tried to copy AWS, didn't fix all the links or think the whole thing through, and just sort of put it out there. I remember we were locked out of our account for like a month for some reason and there was a lot of handoff between different reps, for churn or for some business process reason I wasn't aware of, but it was a pretty jarring experience that made us not want to invest a lot of time there

This is very interesting. My experience at a large company migrating to GCP. The consultants and reps were 10x more eager to help us compared to AWS, and the migration went fairly smooth for most parts

Which aligns with the primary difference between Google and Amazon, customer service. Amazon is as good at service as Google is bad. If my memory serves me correct, Google had outsourced at least some of the GCP customer service to HP.

Customer service isn't wildly different between GCP and AWS. I think it's the finish of the product. AWS has far more finished product and even for smaller services it mostly works as intended. While GCP it is very hit or miss. I have seen so many edge cases in smaller GCP services like composer, old ML platform, dataflow etc.

While GCP is definitely more intuitive. You don't need to create 2 different IAM profile for anything you want to solve. So GCP feels more like Digital Ocean + bunch of services with varying quality.

GCP has a smaller number of services but they tend to be very well designed with excellent taste.

AWS has a bazillion services, often very obviously designed by people without good taste and/or experience.

GCP's incompleteness is super annoying. The general daftness of AWS service design is tiring and somewhat soul-sucking.

I used to think that Google would eventually get there, but lately I've realized that they are totally happy offering an 80% product forever. So for now I'm very reluctantly team AWS.

> GCP has a smaller number of services but they tend to be very well designed with excellent taste.

I would have said the same before using GCP. At first everything goes super smoothly and the experience is definitely better than AWS. Then you start seeing undocumented behaviour again and again. In comparison, it is much rarer to get any undocumented behavior in AWS.

Not my experience. They tend to work well. The problems comes when you want to do something very slightly unusual.

> The problems comes when you want to do something very slightly unusual.

That's exactly what I said. Main services are more or less stable, but there are edge cases in smaller services if you want to do something unusual.

Fully agree with this. I have used GKE at scale as an example. It starts to break down significantly once your cluster reaches a couple hundred nodes. Only way to solve it is to have support resize your controller nodes manually. And as support for anything google is horrible, that task takes multiple hours.

It’s growing because they are slaughtering golden geese at an alarming rate propping up growth

- Youtube ads assault - Endless scrolling on SERP to show more ads as you need to scroll through degrading results. - etc

every surface has increased ad load, every service increased prices. So of course that props up growth.

Ads is in full value extraction mode now, the question is, how sustainable it is, but that’s a question for another quarter

See Facebook the same. You can’t share actual links outside the app anymore which drives engagement as every user clicking on a link now goes back right to being a monthly active

Yes, very noticeable. Rows and rows of ads in Google Search, and I am seeing less and less quality results over time.

Not just ads in YouTube, but a 20% price increase

Interesting news about Ruth Porat transitioning to President and Chief Investment Officer of Alphabet:

> Alphabet and Google CFO Ruth Porat will assume the newly created role of President and Chief Investment Officer of Alphabet and Google, effective September 1, 2023. Ruth will continue to serve as CFO, including leading the company’s 2024 and long-range capital planning processes, while the company searches for and selects her successor.

> In her new role, Ruth will continue to report to Sundar Pichai, Alphabet and Google CEO. > Ruth assumed the role of CFO in May 2015 and is the company’s longest-serving CFO.

> In her new role, Ruth will be responsible for Alphabet’s investments in its Other Bets portfolio, working closely with Sundar, and the company’s investments in countries and communities around the world. Alphabet’s investments span numerous sectors and are engines of economic growth globally. She will also focus on engagement with policymakers and regulators regarding employment, economic opportunity, competitiveness, and infrastructure expansion.

Does this mean cuts a.k.a. "streamlining" incoming for Other Bets? She has been known to be pretty strict about unnecessary spending at Google.

I can't imagine her new role is going to be friendly to Verily.

Why Verily in particular?

Maybe but I think Brin and Page are keen to live forever so Verily is their hedge and research playground in that area. It also serves as a translational testing ground for applied AI innovations in the life sciences.

verily has nothing to do with life extension, that's Calico.

Quarterly numbers from their report:

  - Revenue: $74.6 billion, up 7% YoY
  - Operating income: $21.8 billion
  - Operating margin: 29%
  - Net income: $18.3 billion

How do you play with inflation numbers around the world? 7% YoY does not seem an increase. US official inflation rate is 8% and inflation is calculated based on a basket that is different from inflation based on specific industries.

US inflation YoY was 3% as of the most recent data. https://ycharts.com/indicators/us_consumer_price_index_yoy

4.8% excluding food and energy (as is standard): https://www.bls.gov/news.release/cpi.nr0.htm

Looks like the money machine is still printing! Anyone done deep analysis on how their "other bets" have been performing over time? It feels like a few things have been cut (e.g. Loon) but I haven't been following closely enough to have a good understanding of the full picture

Other Bets Operating Income (Loss): 2022 (1,339), 2023 (813)

Looks like great earnings for probably one of the most undervalued large caps on the market right now (and Amazon).

The company reported EPS of $1.44 vs. Consensus of $1.34 and Revenues of $74.6 Billion vs. $72.85 Consensus.

up 7000 employees despite layoffs.

If you have to let leave some apples without scaring the other ones, doing it while everybody else is doing it is a great chance

I thought this was interesting, so I compared it against their count last quarter (instead of last year) and they are indeed down.

  - In Q4 2022 [1] they had 190,234 employees.
  - In Q1 2023 [2] they had 190,711 employees.
  - This quarter Q2, they have 181,798 employees.

So they are down almost 9000 employees from last quarter, close to their 12K layoff number.

[1] https://abc.xyz/assets/c4/d3/fb142c0f4a78a278d96ad5597ad9/20... [2] https://abc.xyz/assets/a7/5b/9e5ae0364b12b4c883f3cf748226/go...

they could fire half their workforce and they'd be back to 2019 numbers

https://www.statista.com/statistics/273744/number-of-full-ti...

It's been about 6 months since the Google CEO blogged about how they needed to lay of 12,000 people to brace for the "difficult economic cycle". Meanwhile revenue is up 7% YoY.

https://blog.google/inside-google/message-ceo/january-update...

Looks like they weren’t really needed to add to that revenue.

If that is true then it means they hired more people than they needed and then instead of admitting it they gave some nonsense about macro economics

They hired more really good people to leave fewer really good candidates to the competition. Maybe it made sense. They could not publicly declare this though.

Maybe they hired more than they needed because of the shifting macro environment? The number of employees you need depends on what you want to accomplish which depends on interest rates, consumer spending, and tons of other macro factors.

They always hired more than needed, even now. I think it's a common knowledge. Most of the time it worked out because google's stock always grew, but yeah if conditions become bad they wouldn't need to think a lot before laying of 20-30% of the workforce.

I am friends of a long time, Fortune 50 CFO. Well, he was about 16 years ago before one large company went public and he made about $40M in a single day and decided he wanted to do something new (can't blame him).

He told me repeatedly that companies that lay off the moment a bad quarter appears - or if they lay off employees if they even think the market will slow down - are always terribly run companies.

He said it was a clear indicator of poor internal planning and forecasting, that any company who suddenly needed to shed 5,000 or 10,000 employees on one bad quarter was was one to avoid.

This can be true and we can also recognize that the pandemic probably made everyone’s internal forecasts bad. These tech companies have some smart people working there, I would bet their projections are pretty world class.

I've worked there, it's not any different than the rest of the business world.

How so? The motivation and competency and ownership is the same as at say Comcast?

The comcast is bad meme is interesting but what specifics do you have that it is worse at business than Google?

I am no fan of comcast and pay a premium to avoid them since I disagree with bandwidth quotas. But both Google and Comcast have bad customer service and some terrible products yet both have respective and respectable monopolies. YouTube is terrible and amazing at the same time and more teens watch it than cable tv. Yet there is awful content and no good way to screen it from children besides denying access entirely.

Google also arguably missed ChatGPT and TikTok as macro product trends. Previously, Google failed to capitalize on social networking.

I am not sure how you forecast or predict those things in any way that guarantees success. And even when you see the future to pursue and build that core competency. I don't think large companies can move fast enough because by the time you notice the competition they've already established a network effect.

And network effects seem to be the main moat. It remains to be seen if Threads really kills Twitter. It's not even clear that Musk can kill it unintentionally, i.e. the business succeeds despite its management.

A company the size of Alphabet "shedding" 5-10K employees is a notable but still minor adjustment in the grand scheme of things.

No, it isn't. Details matter, and if they hiring even 1% too many people, then what else are they getting wrong by 1%? And how many of those things?

The key is that they wouldn't know either. It could be a slow 'death by a thousand cuts' and they wouldn't even know they were bleeding.

This is why companies, even massive companies, generally die eventually. There are very few that last even 50 years without eventually being acquired or bankrupt. And how does it start? Little mistakes, lack of focus, everywhere.

> then what else are they getting wrong by 1%?

They could get 1000 things wrong by 1% and they're still 1% wrong. You can't really add them up.

Touché.

I mean, everyone knows that PPC is 95% of Google's revenue business. They could probably eject 80% of the company and the revenue trajectory would not change.

If Twitter's (or X?) performance is anything to go off of, that's probably not true.

Right. Twitter was such a money spinner before.

Comparatively more than it is now. It's a no-true-scotsman to exclude it.

All of Twitter's staff were supporting the core product. How many Google employees support no part of the Google Ads stack or a product with Google Ads revenue?

Employee count actually increased accounting for layoffs, so that's not it for sure. Looks like they did need or at least thought they needed that many.

Looks like the number of employees increased from 174,014 to 181,798. This includes the 12,000 laid-off employees.

I just added this comment on another reply further down below, copy-pasting here, the numbers compare last year Q2 to this year Q2, if you compare it to last quarter they are actually down in count.

-------------------------------

I thought this was interesting, so I compared it against their count last quarter (instead of last year) and they are indeed down.

  - In Q4 2022 [1] they had 190,234 employees.
  - In Q1 2023 [2] they had 190,711 employees.
  - This quarter Q2, they have 181,798 employees.

So they are down almost 9000 employees from last quarter, close to their 12K layoff number.

[1] https://abc.xyz/assets/c4/d3/fb142c0f4a78a278d96ad5597ad9/20... [2] https://abc.xyz/assets/a7/5b/9e5ae0364b12b4c883f3cf748226/go...

The layoffs aren't finished.

This is interesting because it then is such a sharp U-turn since last Q2. "Keep hiring!" -> "Layoffs!" with very little pause in between?

Yeah pretty much all major tech companies did this exact same U-turn.

Yeah, but seeing the sharpness - compared to "ok let's slow down... ok ... shoot, we probably actually need to trim some headcount" - really highlights the follow-the-herd, here's-a-handy-excuse, done-for-PR nature of them vs actually responding to revenue trends or such.

So inflation adjusted they’ve experienced no growth.

Inflation is not at 7%.

Average global inflation was 8.7% last year, and in the US it peaked at 9.1% in June 2022. The calculation is obviously a lot more complex considering all the different jurisdictions Google makes revenue from and all the different kinds of measures, but adding it all up it won't be too far from the 7% number.

the inflation rate at June 2022 is not what you should use to normalize the growth of a company, YoY, at June 2023. You're off by a year. The US YoY inflation rate for this period was around 3% going off memory.

Idk what you define as "average" global inflation. I don't really agree that it's the correct term for what is largely an american company even if it does international business.

Inflation doesn't jump from 9% to 3% overnight but goes down day-by-day. It is only at 3% now, so it spent the majority of the last year at >6%. And Google makes 55% of its revenue from outside the US, so it makes sense to consider global inflation in the calculation not just US.

3% is the average rate over the past 12 months. "it spent the majority of last year at > 6%" is not a helpful metric because those past numbers are average of months prior to June 2022.

We are tracking growth from June 2022->June 2023. So what we care about is the inflation from June 2022->June 2023. Which is 3%.

It doesn't matter what inflation is in other countries if the vast majority of Google's earnings are in dollars.

> if the vast majority of Google's earnings are in dollars

They aren't.

True, but the word "currency" pops up 53 times in this report [1], mostly in regards to managing "foreign exchange risk", so they seem to be taking deliberate measures to stay on the inflation trajectory of the dollar.

[1]: https://www.sec.gov/Archives/edgar/data/1652044/000165204423...

Every single number in this report is in dollars.

It is in many parts of the world.

There are many parts of the world though.

People make mistakes and economic conditions change...

Weren't most of the layoffs in big tech done to shed positions related to recruiting or kill off floundering projects in R&D? If so, revenue is the wrong number to be looking at.

no, very much not the case.

It depends. Plenty of companies did big vertical cuts, affecting every team, department, job function and level equally.

Microsoft spun the wheel and laid off all sorts of people. Not sure of the rhyme or reason.

https://en.wikipedia.org/wiki/The_Time_Machine

... A work of future history and speculative evolution, Time Machine is interpreted in modern times as a commentary on the increasing inequality and class divisions of Wells' era, which he projects as giving rise to two separate human species: the fair, childlike Eloi, and the savage, simian Morlocks, distant descendants of the contemporary upper and lower classes respectively ...

... Deducing that the Morlocks have taken his time machine, he explores their tunnels, learning that due to a lack of any other means of sustenance, they feed on the Eloi. The Traveller theorizes that intelligence is the result of and response to danger; with no real challenges facing the Eloi, they have lost the spirit, intelligence, and physical fitness of humanity at its peak. ...

After hours trading: GOOG up 6.8% to $131/share

> Google Services includes products and services such as ads, Android, Chrome, hardware, Google Maps, Google Play, Search, and YouTube. Google Services generates revenues primarily from advertising; sales of apps and in-app purchases, and hardware; and fees received for subscription-based products such as YouTube Premium and YouTube TV.

Interesting how they bundle chrome, android, and pixel with various web services such as search, yt, maps

Kind of odd considering pixel is a completely different kind of product. It does make sense because there's likely synergy between first group and second group, but also it makes me wonder how they calculate the profitability of a product like pixel

Also it makes me wonder whether the recent proposal of the chrome integrity API will positively or negatively effect these numbers long term given sufficient backlash

I ran some google ads lately and they were very ineffective. I feel like my money was going down the drain. Perhaps that's what s driving their higher revenue?

Ineffective ads wouldn't drive higher revenue. It would be the opposite.

Ineffective ads require higher spend than the effective ones.

At 100% effectiveness (100% conversion rate) you only need to buy N ads to target your sale target of N.

Less effective ads mean you need to buy more of them

But that also means they’d be cheaper.

Advertisers don’t buy ads unless they have measurable ROI. It doesn’t matter if it’s 1000 impressions for $1 for 1 conversion, or 10 video views for $1 for 1 conversion, or a million impressions for $1 for 1 conversion. It’s just money spent vs money gained.

Marketers should spend every extra dollar on the ad platform with the lowest customer acquisition cost, so if Google ads are more ineffective compared to others, then its customer will go somewhere else and hence it will have less revenue.

During the social media ad pause of the 2020, we had multiple companies admit in earnings calls and elsewhere that pausing advertisement had no impact on their sales at all.

The extreme end of that spectrum is Tesla with no advertising whatsoever.

This kind of observation makes me pause and cause some reflection on ads in general.

Nit: Tesla has word of mouth advertising, which works for automobiles when you have brand loyalty.

Do they pay for it?

Well they have to pay some engineers occasionally to work on flamethrowers.

Yes - https://www.tesla.com/support/refer-and-earn

Even at hypothetical 100% effectiveness, you're still going to be spending the same budget to match variable ROI on CPCs and CPA.

I was tempted to bet against them since every Google product seems to be getting worse- felt like getting blood from a stone.

There's a delay between the enshittification and the stock peak, we're in that window now.

So you think there's still a peak; are you buying lots of GOOG right now?

Yep, along with 499 other large cap US companies.

Reply on news.ycombinator.com