zlacker

[parent] [thread] 74 comments
1. bko+(OP)[view] [source] 2025-05-06 12:46:45
Incredible timeline to a $3B exit

> Windsurf began in 2021 as Exafunction, founded by MIT graduates Varun Mohan and Douglas Chen. The company initially focused on GPU optimization before pivoting to AI-assisted coding tools, launching Codeium, which later evolved into Windsurf.

> Series B (January 2024): $65 million at a $500 million valuation.

> Series C (September 2024): $150 million, led by General Catalyst, at a $1.3 billion valuation.

> May 2025: $3 billion acquisition from OpenAI

I wonder how much of the value is really from the model or the tooling around it. They all use the same models (mostly Claude, others have been horrible and buggy in my experience). Even co-pilot agent mode now uses Claude. The editor has their own LLM (?) that does the apply since LLMs often return snippets. They work well enough on Cursor. And then you have the auto-complete, which I think is their own model as well.

But the main value from me is from the agent mode and 95% of the value is the underlying model. The other stuff could be more or less a VS Code plugin. The other benefit is the fixed pricing. I have no idea how much 500 calls cost if I were to use the API, but I expect they're probably losing money.

replies(13): >>bfeynm+M2 >>dist-e+H6 >>morale+y21 >>rpgbr+0f1 >>mvkel+VS1 >>sagarp+TY1 >>SwtCyb+Je2 >>froh+Mi2 >>Iolaum+3y2 >>ulfw+xC2 >>qwytw+RC2 >>meerit+uF2 >>bcx+sY2
2. bfeynm+M2[view] [source] 2025-05-06 13:06:33
>>bko+(OP)
talented and smart folks for sure but can't not notice how much luck it is especially because its like 100% just better models. Windsurf raised a ton of money and then said they pivoted which they had millions raised to just do something completely different that likely wouldn't have been easier to raise for. Even in an interview with the cursor founder he kind of dumbly rambles that they launched and then basically lost a ton of traction until GPT4 came out. They have some core features like autocomplete but I'm struggling to see vision other than getting training data for iterative dev is a partial moat compared to just seeing commits and final code bases.
replies(1): >>sigmoi+hV1
3. dist-e+H6[view] [source] 2025-05-06 13:28:46
>>bko+(OP)
> I wonder how much of the value is really from the model

> The other stuff could be more or less a VS Code plugin

The other stuff would take a team 6 months to implement. This is where the valuation comes from. Time to market, they are there TODAY.

replies(2): >>suppor+Uc1 >>pclmul+eQ1
4. morale+y21[view] [source] 2025-05-06 18:57:14
>>bko+(OP)
The right time and the right place, plus they did the work, ofc; but I'm sure 80% of this site has worked as hard as, or even more, than what it takes to clone VSCode.

I'm jelly. Very rarely you see in history someone lucky enough to be riding the absolute top of the wave. Even OpenAI took about decade to cook their breakthrough product.

◧◩
5. suppor+Uc1[view] [source] [discussion] 2025-05-06 20:09:55
>>dist-e+H6
6 months of anyone's time is not worth 3 billion dollars.
replies(2): >>quanta+ZU1 >>docker+r12
6. rpgbr+0f1[view] [source] 2025-05-06 20:21:47
>>bko+(OP)
It's a bubble about to pop. That's where the value is coming from.
replies(3): >>ignora+FP1 >>throw2+H32 >>madduc+zs2
◧◩
7. ignora+FP1[view] [source] [discussion] 2025-05-07 01:58:28
>>rpgbr+0f1
Bubble or not, given the exit, Windsurf's (Codeium) focus on enterprise sales motion has been rewarded rather handsomely: https://research.contrary.com/company/windsurf / mirror: https://archive.vn/ThWNz
replies(1): >>alexch+QQ1
◧◩
8. pclmul+eQ1[view] [source] [discussion] 2025-05-07 02:07:00
>>dist-e+H6
As usual, HN misses the point. The customer list was probably worth about $2.99 billion, and the engineering work about $10 million.
replies(3): >>mjirv+qS1 >>mplewi+KS1 >>pqtyw+VD2
◧◩◪
9. alexch+QQ1[view] [source] [discussion] 2025-05-07 02:14:04
>>ignora+FP1
Yeah, in the recent Lightcone Podcast episode, Varun was talking about how they have a lean eng team but large sales org. I thought that was super interesting for a dev tool since I was expecting a dev tool to involve bottom-up sales to the dev instead of top-down sales to a leader like a CTO or VP of Eng
replies(2): >>conart+Sb2 >>tracer+1Z3
◧◩◪
10. mjirv+qS1[view] [source] [discussion] 2025-05-07 02:35:34
>>pclmul+eQ1
Is OpenAI having trouble acquiring enterprise customers?
◧◩◪
11. mplewi+KS1[view] [source] [discussion] 2025-05-07 02:39:42
>>pclmul+eQ1
What makes you think $3B worth of customers were committed to Windsurf at all, much less in a sticky/exclusive way?
12. mvkel+VS1[view] [source] 2025-05-07 02:40:48
>>bko+(OP)
The value is in the prompts being sent to OpenAI. Massive training depository.

Only thing better would be a social network, which supposedly they're working on.

replies(1): >>buffer+CX1
◧◩◪
13. quanta+ZU1[view] [source] [discussion] 2025-05-07 03:03:42
>>suppor+Uc1
That was my thoughts too. No text editor is worth $3B, and probably not even VSCode is. So I think this deal was about buying more customers/users and buying "relevance". OpenAI lost it's monopoly and they're worried they might become irrelevant so they basically just purchased something popular to remain relevant.
◧◩
14. sigmoi+hV1[view] [source] [discussion] 2025-05-07 03:06:52
>>bfeynm+M2
Training data is almost certainly their main reason for this acquisition. Users themselves and the models they use don't really matter. What matters is their interactions with the models. Especially if you're trying to build coding agents that will be marketed to companies for $10k a month. OpenAI is going for the industry B2B opportunity here, not consumers or end users.
replies(1): >>thomas+2F2
◧◩
15. buffer+CX1[view] [source] [discussion] 2025-05-07 03:34:26
>>mvkel+VS1
But OpenAI already knows every single prompt sent to its models. They don't need to buy Windsurf for that.
replies(3): >>Mtinie+f12 >>Nitpic+t22 >>mvkel+K22
16. sagarp+TY1[view] [source] 2025-05-07 03:51:30
>>bko+(OP)
I’ve been a WindSurf customer since day one. It was my first true AI agentic experience.

[Dev mode] While working on Alembic migrations I broke one of my migration files. After an hour of manual debugging I handed the task to WindSurf. It methodically checked every config file, applied the migrations one by one, and narrowed the issue to a single file. It rewrote the migration, verified the fix, wrote tests, ensured everything passed, and opened a PR. I reviewed it and it worked flawlessly.

Regarding the acquisition I don’t understand why OAI would pay $3 B. The team is strong, they have lots of data, and the agentic flow is great, but all of that feels commoditized.

Claude Code launched two months ago and I prefer it to WindSurf, Cursor, and Aider. Augment Code also ranks above WindSurf for me.

If I were in Sam’s place I would have doubled or tripled down on Codex CLI. Just my 2 cents.

replies(9): >>advent+OZ1 >>moeadh+X22 >>anonzz+vb2 >>bingem+ul2 >>jjalle+qJ2 >>dayjah+HL2 >>boring+VM2 >>neets+Vw6 >>shift8+Lc7
◧◩
17. advent+OZ1[view] [source] [discussion] 2025-05-07 04:04:00
>>sagarp+TY1
They're paying $3 billion because money is hyper plentiful for OpenAI at present. Basically because they can. Money isn't their problem right now, it's not a scarce resource (maybe it will be in the future of course). They're trying to capture and lock-in, so as the hurdles and regulations go up they're one of the huge winners left standing.

Try replacing Uber today, it's impossible. Nobody is going to give you billions of dollars to try to do it. It'd be an absolute nightmare to attempt it.

replies(1): >>mike_h+bl2
◧◩◪
18. Mtinie+f12[view] [source] [discussion] 2025-05-07 04:19:06
>>buffer+CX1
Now they also get to see what is sent to Anthropic, Google, DeepSeek, etc., and what is returned. At scale, for a prime area of concern.
replies(2): >>buffer+v22 >>r0b05+I22
◧◩◪
19. docker+r12[view] [source] [discussion] 2025-05-07 04:22:17
>>suppor+Uc1
it's not just one person though
◧◩◪
20. Nitpic+t22[view] [source] [discussion] 2025-05-07 04:37:45
>>buffer+CX1
There's much more to be gained if you also have the client side of those interaction. You can get signals from "accepted" completions/plans/etc, number of edits made to those completions, how users use context, what was passed in context from a code base, and so on.

And that's just on their models. They'd also get (at the very least) signals on their direct competition, if not straight up prompts+completions as well.

◧◩◪◨
21. buffer+v22[view] [source] [discussion] 2025-05-07 04:37:52
>>Mtinie+f12
Reread the comment I replied to. It has nothing to do with Anthropic or Google or DeepSeek.
replies(1): >>Mtinie+cT9
◧◩◪◨
22. r0b05+I22[view] [source] [discussion] 2025-05-07 04:40:04
>>Mtinie+f12
That's valuable user and competitor data actually.
◧◩◪
23. mvkel+K22[view] [source] [discussion] 2025-05-07 04:40:34
>>buffer+CX1
That's a bit like saying having access to Google is as good as being Google.

All they really see as a model provider is little fragments of the picture, like trying to reconstruct the Mona Lisa by knowing which paint swatches Leonardo used.

In other words, they only saw whatever Windsurf sent as context with a "fix the bugs" prompt stapled to it.

By owning Windsurf, they see the entire source code of what's being built, all the time, plus how the model is interacting with it.

There's a massive amount of value in what happens client-side, and behind the scenes. The "director's cut" of context.

Huge difference.

replies(1): >>pqtyw+pD2
◧◩
24. moeadh+X22[view] [source] [discussion] 2025-05-07 04:44:33
>>sagarp+TY1
sometimes companies are acquired for things the public has not yet seen.
replies(2): >>johnta+v52 >>arrows+dI2
◧◩
25. throw2+H32[view] [source] [discussion] 2025-05-07 04:54:13
>>rpgbr+0f1
5 years ago if you said coding tools would be worth in the billions in value it would of been surprising to most people. Dev Tools were the thing you could never get a company to buy for you or were just free for most people. Interesting times.
replies(1): >>mike_h+ol2
◧◩◪
26. johnta+v52[view] [source] [discussion] 2025-05-07 05:23:09
>>moeadh+X22
Companies are acquired for customer base, ARPU, and growth. Same criteria as when when raising funding.
replies(2): >>jahews+pa2 >>jilles+Yg2
◧◩◪◨
27. jahews+pa2[view] [source] [discussion] 2025-05-07 06:30:40
>>johnta+v52
Those are the best reasons, but companies are also acquired for marketing, hype, to relieve a sense of fear, to curry favour, etc…
◧◩
28. anonzz+vb2[view] [source] [discussion] 2025-05-07 06:44:57
>>sagarp+TY1
> WindSurf, Cursor, and Aider. Augment Code also ranks above WindSurf for me.

Bring on (a lot) more competition! I am waiting for the point where "Simple Pricing" (Augment Code has that on the pricing page) means fixed pricing; Simple is NOT '600 messages included' because it's impossible to know what the ROI from those 600 messages is, so it's very far from 'Simple' (many of those prompts will deliver nothing or, worse, having to rollback because the agent produced garbage). I know it's not sustainable, but the only thing that will keep me not jumping from one to the other, signing up with different emails, trials, coupons etc is if they will lose the restrictions on usage. They will, as they have to compete and it's worth it seeing this acquisition; losing 10s of millions a month to get/keep people and getting nice growth is what they do to get the billions. So bring it on!

◧◩◪◨
29. conart+Sb2[view] [source] [discussion] 2025-05-07 06:50:32
>>alexch+QQ1
It will be super interesting to see how they do against the inverse: an engineering-focused company that wants to win devs from the bottom up
replies(1): >>alexch+q43
30. SwtCyb+Je2[view] [source] 2025-05-07 07:28:47
>>bko+(OP)
Totally agree - a lot of the "magic" still feels like it boils down to whoever has the best underlying model
◧◩◪◨
31. jilles+Yg2[view] [source] [discussion] 2025-05-07 07:59:02
>>johnta+v52
I don't think that's the case here. Windsurf wasn't leading the agentic coding market. They were doing a decent job but others are bigger. Cursor has the brand recognition and Claude is getting a lot of recognition too. MS has github copilot which is still a good brand and Google has been catching up with Gemini.

OpenAI has a new thing called codex but it isn't very good yet. I tried it and it's super flaky. Lot's of errors and it gets stuck when that happens. OpenAI needs something good urgently because agentic coding is the key AI feature right now and the blue print for non coding agentic solutions later. Cursor is probably too expensive currently and windsurf looks like their models are a bit better.

So, OpenAI gains something they don't have: a credible developer option with an active user base and some core IP in the form of training data and know how as well as custom models that they can fold into openai.

3 billion is a lot but not if you consider that world + dog in the enterprise world will be spending big time on AI subscriptions for their developers. This stops being optional in 2025. Millions of developers will be on paid subscriptions permanently very soon. If you start a new job you can expect to get a laptop and a paid subscription to whatever is the agentic coding tool of choice in your new company.

OpenAI wants double digit percentages of that revenue. 1M users paying something like 50$/month would amount to 600M revenue per year. I think the prices will go up and the amount of active users as well. Reason: as these tools are getting better they start saving non trivial amounts of engineering time. At that point you have to value the tool in terms of developer cost. Not 1 to 1. But it's worth a sizable portion of that.

I work in a small startup as the CTO. This is an no-brainer for us. We're cash strapped so we only spend on important things. This would be one of those things. We're doing things I previously would have needed to expand the team for because I would have had no capacity to do those things in the current team. So, in terms of value for money spending on these tools is easy to justify.

I get lots of people are skeptical about AI stuff here. But I would say that a lot of those people suffer from a short term focus and bias. Three years ago none of this stuff existed. Now it's a multi billion$ market that is set to grow rapidly. Stuff is getting better at a very rapid pace. Just stating facts here. 3 billion is a bargain if Openai can make this acquisition work for them. They are buying time to market here. They don't have a year to figure it out. In a year or so this market will be carved up and locked into hard to change year long SAAS contracts. At that point getting people to switch tools will get harder and harder.

replies(2): >>pantul+ds2 >>justan+ns2
32. froh+Mi2[view] [source] 2025-05-07 08:19:45
>>bko+(OP)
The value is the team and their thinking and the customer base.
replies(1): >>pqtyw+BD2
◧◩◪
33. mike_h+bl2[view] [source] [discussion] 2025-05-07 08:48:21
>>advent+OZ1
Uber has already been replaced, at least in some parts of the world. We recently went on holiday to Malta and on check-in the hotel staff told us not to bother with Uber, Bolt worked way better and had more drivers (Bolt is a European Uber competitor based in Estonia).

So we signed up for Bolt and sure enough drivers were plentiful, the app worked great and there was no downside over Uber. I'll certainly be trying it again in future in other markets.

The reason Uber invested in self-driving cars for years is that otherwise they have no sustainable edge. It's just a taxi company, which is a low margin business. People who can make slick mobile apps are plentiful and it takes a minute or less to sign up for a new service. Uber grew to its current size by buying market share using investors dollars, which was always a time-limited strategy. Once they started having to turn a profit prices rose and their edge over their competitors was lost.

replies(1): >>petese+pm2
◧◩◪
34. mike_h+ol2[view] [source] [discussion] 2025-05-07 08:49:56
>>throw2+H32
Dev tools are still very hard to sell (I know, I have a dev tools company). Claude Code, Aider and Codex are given away for free. What people are buying is access to proprietary general purpose models.
◧◩
35. bingem+ul2[view] [source] [discussion] 2025-05-07 08:50:56
>>sagarp+TY1
Is there a writeup or a recording. Would be nice to follow through
◧◩◪◨
36. petese+pm2[view] [source] [discussion] 2025-05-07 09:01:31
>>mike_h+bl2
Uber feels like such an apt comparison to OpenAI to me. The service they provide is obviously going to be absolutely huge, but no guarantees at all that they’ll win it or be last man standing. I don’t see a world in which generative AI doesn’t continue to be a massive disrupting force, but no particular reason to think Anthropic or OpenAI will still be independent entities in a few years.

I’m even more bearish on Uber than I used to be, as someone who’s used Grab and Careem and Bolt extensively, and seen Uber have to beat a retreat from SE Asia. If their more nimble competition get a foothold in the US they’re toast.

replies(1): >>saalwe+ZT2
◧◩◪◨⬒
37. pantul+ds2[view] [source] [discussion] 2025-05-07 10:06:35
>>jilles+Yg2
> OpenAI has a new thing called codex but it isn't very good yet. I tried it and it's super flaky. Lot's of errors and it gets stuck when that happens.

I agree with this, not sure the experience of everyone else but I felt like Claude Code is more useful.

Meanwhile, I'm keeping tabs on Aider and open-codex, what other options are there?

replies(1): >>therea+ri3
◧◩◪◨⬒
38. justan+ns2[view] [source] [discussion] 2025-05-07 10:08:06
>>jilles+Yg2
I wanna pick your brain a lil. Are you saying agentic AI has helped you replace devs that you would otherwise need for your startup?

In your opinion, what should one do/learn to get a SDE/related gig now? What do you/other companies look for?

replies(2): >>jilles+Fz2 >>indymi+UR3
◧◩
39. madduc+zs2[view] [source] [discussion] 2025-05-07 10:09:41
>>rpgbr+0f1
It's all about stocks
40. Iolaum+3y2[view] [source] 2025-05-07 11:14:49
>>bko+(OP)
I think part of the value is customer acquisition rather than product.
◧◩◪◨⬒⬓
41. jilles+Fz2[view] [source] [discussion] 2025-05-07 11:35:15
>>justan+ns2
Not replace but it allows me to scale what we do for things we previously would have dropped because it would require growing the team, which we can't really afford. It's a case of getting a bit more out of developers in terms of quantity and scope (mostly this) of what they do. Not a full developer but enough for it to be meaningful. But it's not nothing either. Worth paying for. AI is a lot cheaper than a developer is so I don't need to replace my developers. I prefer people that are multi disciplinary and able to pick up new skills as they are needed. Agentic AIs are good for that because they give you enough to work with that you can get productive with whatever you need to wrap your head around in little to no time.

Companies can be a bit slow to update their hiring processes to their needs. But good developers should be ahead of the curve in any case. For this, just be proficient with the tools.

Be ready for the inevitable interview question "so, AI ... explain me how you are using it and what you are doing with it?". Much easier to answer that question if you have some meaningful time of routinely using this stuff behind you and can articulate what works and doesn't work for you.

And if they don't ask, that's actually a great question to ask back if you get the opportunity "I've been using agentic tooling, how are you guys using that a <company name>? Also I would like a subscription to <my favorite AI tool> if I work for you". Stuff like that makes you stand out as ambitious and interested in the future. There are of course going to be places that maybe don't like that. But then ask yourself whether you'd want to work there. So, either way, you learn something.

replies(1): >>justan+tB2
◧◩◪◨⬒⬓⬔
42. justan+tB2[view] [source] [discussion] 2025-05-07 11:50:06
>>jilles+Fz2
Thank you for responding :)

So, would you consider it a bad idea to get into web dev, more specifically backend and infra?

Do you think using LLMs can accelerate learning software dev and programming skills?

replies(1): >>jilles+pI2
43. ulfw+xC2[view] [source] 2025-05-07 11:58:58
>>bko+(OP)
It's a beautiful world where you'll only put a little over 220 MM in and get 3000MM out mere months later.
44. qwytw+RC2[view] [source] 2025-05-07 12:00:46
>>bko+(OP)
> Incredible timeline to a $3B exit

dot-com vibes. Maybe not quite the same as Pets.com but still...

◧◩◪◨
45. pqtyw+pD2[view] [source] [discussion] 2025-05-07 12:05:04
>>mvkel+K22
So just put up together a comparable VS Code based AI IDE in a couple of months and bundle it together with the ChatGPT subscription? They'd get loads of users very fast..
replies(1): >>sanxiy+AG2
◧◩
46. pqtyw+BD2[view] [source] [discussion] 2025-05-07 12:06:22
>>froh+Mi2
How is it different to a bunch of other apps which and tools which offer more or less exactly the same?
◧◩◪
47. pqtyw+VD2[view] [source] [discussion] 2025-05-07 12:08:59
>>pclmul+eQ1
> customer list

How many customers do they have? At $30 per month it would take forever to pay off even with a lot of growth.

Open AI could release an equivalent VS Code clone and make it entirely free and it would still be a lot cheaper than $3 billion.

◧◩◪
48. thomas+2F2[view] [source] [discussion] 2025-05-07 12:19:01
>>sigmoi+hV1
But aren't they getting this data already at a much larger scale? GPT is still one of the backbones in many coding assistants, even Windsurf.
replies(1): >>sigmoi+EF7
49. meerit+uF2[view] [source] 2025-05-07 12:21:43
>>bko+(OP)
Visual Studio Code Agent Mode uses whatever model you tell it to use.
◧◩◪◨⬒
50. sanxiy+AG2[view] [source] [discussion] 2025-05-07 12:30:12
>>pqtyw+pD2
I think it is exactly this. There is no doubt whatsoever OpenAI can do this, but they decided not to. The reason, I think, is that they don't want to be a couple of months late. In other words, they spent $3B to save a couple of months.
◧◩◪
51. arrows+dI2[view] [source] [discussion] 2025-05-07 12:42:57
>>moeadh+X22
In this case, maybe it's an acqui-hire?
replies(1): >>lintco+fi3
◧◩◪◨⬒⬓⬔⧯
52. jilles+pI2[view] [source] [discussion] 2025-05-07 12:44:44
>>justan+tB2
I would look forward to the next 20 years and not backward to the last 20.

The whole frontend/backend distinction did not really exist until the web. And infrastructure is definitely something that should be automated far more than it currently is. If it needs babysitting by a team of devops, you just created a lot of work rather than automating/solving it. Tedious and repetitive. It has "AI will make this a lot easier" written all over it.

So, just be ready for the ambition level to be raised for developers. Learn to build the whole system, not just bits and pieces of the system. Lean on AI to get stuff done and figure things out. It's all just code. None of it is really that hard. But it can be a lot of work if you do all of it manually.

And let's be honest, agentic tools are showing promise and great progress but they are nowhere close to independently working on existing code bases. That's not how I use them. But they are great for problem solving, debugging, prototyping, exploring some new languages and APIs, and generally taking care of more tedious coding tasks.

◧◩
53. jjalle+qJ2[view] [source] [discussion] 2025-05-07 12:50:19
>>sagarp+TY1
Because they have users and OpenAI has seen the massive drop off in coding usage since Claude Code came out. My personal Chatgpt decline is at least 99%. It’s also 1% of their current market value. So not really a big deal.
◧◩
54. dayjah+HL2[view] [source] [discussion] 2025-05-07 13:06:08
>>sagarp+TY1
I was involved in an M&A once; my role was to evaluate the technology and determine how long it would take us to build a competitive product. If it was less than some X then we’d build it, greater than X and we’d buy. The function for X was not clear to me from my perspective; it had legal fees involved, etc.

The person leading M&A said an intangible aspect of the price is what it does to the adjacent market. If the leading product A is valued during a raise at $Y, and you buy the next best product B at 1/10 that, you cause future issues with raises for A.

Might this be an attempt to clip Cursors wings?

replies(2): >>pc86+4X2 >>foobar+qw3
◧◩
55. boring+VM2[view] [source] [discussion] 2025-05-07 13:12:40
>>sagarp+TY1
If he's (Sam) trading equity on a grossly inflated OpenAI for the acquisition then he's likely not paying real money for the company and thus he is expanding his roadmap for cheap.
◧◩◪◨⬒
56. saalwe+ZT2[view] [source] [discussion] 2025-05-07 13:46:35
>>petese+pm2
I feel like the value-proposition of Uber was three-fold.

1. Solving a pain-point of many people re: hailing a cab, via an app that works everywhere.

2. Using VC funds to (initially) pay drivers more than you, the customer, were paying them.

3. Ignoring local regulations and passing the savings/convenience on to you.

1 is nice but I don't think they established much of a moat (both drivers and customers are willing to use multiple apps); 2 isn't sustainable in the long-term, and they failed to leverage 3 to establishing a permanent right to operate as they had been in most markets.

I think this makes Uber an even more interesting benchmark for other unicorns, since besides "solving a real problem without establishing a moat" they are also often burning through VC cash to prop up their business model while ignoring some laws which they may not be able to get away with ignoring long-term.

replies(1): >>datadr+X73
◧◩◪
57. pc86+4X2[view] [source] [discussion] 2025-05-07 14:01:50
>>dayjah+HL2
That's a really interesting thought, I'd love to get involved in software PE/M&A on the technical analysis side but I don't have the academic pedigree for it (it seems every shop that does this work is 90% Ivy and Ivy-adjacent universities and FAANG-level work history).

So if I'm understanding your point then part of the value in paying $3B for Windsurf is that it acts as a pricing anchor on future raises (and presumably acquisitions as well) for competing products? So Cursor is less likely to raise at a $30B valuation if Windsurf is 95% as good and just sold for 1/10 that.

replies(2): >>whatsh+2J3 >>dayjah+cM4
58. bcx+sY2[view] [source] 2025-05-07 14:08:14
>>bko+(OP)
Incredible timeline - also helpful to understand the OpenAI side.

1) OpenAI is valued at 300B (as of March 31st) https://openai.com/index/march-funding-updates/

2) OpenAI recently raised 40B from SoftBank and others.

3) Windsurf is getting roughly 1% of OpenAI's valuation.

OpenAI needs to keep moving fast to outpace MS, Google, and others -- and I think we can all agree that agentic coding is a major trend -- that is likely to keep growing really fast -- and super high leverage in that the folks doing the coding are well paid -- and more likely to be early adopters than any other field. (e.g. if openAI wants a fast way to grow beyond $20-$200/month, owning a tool like windsurf is a good move)

Some folks have been speculating the cash/equity split. I'd be confident whatever number they arrived at de-risks things for windsurf, and preserves the right amount of cash on hand for openAI.

Even if OpenAI is burning 10-20B a year, with the recent raise would buy them between 1-2 years, and given the pace of AI development that's a pretty long time.

◧◩◪◨⬒
59. alexch+q43[view] [source] [discussion] 2025-05-07 14:36:21
>>conart+Sb2
100%. You're referring to Cursor?
◧◩◪◨⬒⬓
60. datadr+X73[view] [source] [discussion] 2025-05-07 14:51:56
>>saalwe+ZT2
1 especially is a social function. Having to have a million different apps is terrible, but if there is too much competition for drivers it's inevitable to churn through apps because of marketplace pricing and rent seeking on all sides
◧◩◪◨
61. lintco+fi3[view] [source] [discussion] 2025-05-07 15:40:00
>>arrows+dI2
doubtful imo
◧◩◪◨⬒⬓
62. therea+ri3[view] [source] [discussion] 2025-05-07 15:41:06
>>pantul+ds2
Thanks for mentioning open-codex. Did not notice that there is a codex fork which is open to other models (update: totally missed that original codex allows that too now). How do you like it? Especially in comparison to Claude Code?
replies(1): >>pantul+ni5
◧◩◪
63. foobar+qw3[view] [source] [discussion] 2025-05-07 16:53:43
>>dayjah+HL2
I would also think that a critical component of X there would be the opportunity cost of time spent on building in-house while competition chugs along.
replies(1): >>dayjah+CM4
◧◩◪◨
64. whatsh+2J3[view] [source] [discussion] 2025-05-07 18:01:14
>>pc86+4X2
The insanity of it all is that these companies are worth about $3M.
◧◩◪◨⬒⬓
65. indymi+UR3[view] [source] [discussion] 2025-05-07 18:55:00
>>justan+ns2
> I wanna pick your brain a lil. Are you saying agentic AI has helped you replace devs that you would otherwise need for your startup?

I need fewer devs to get more work done... but interestingly it has put a premium on experience because a lot of the "human work" is debugging and fixing where the LLM missed the mark. So less headcount, higher skill required.

replies(1): >>justan+l5a
◧◩◪◨
66. tracer+1Z3[view] [source] [discussion] 2025-05-07 19:35:47
>>alexch+QQ1
thats how you spend a lot of time making something great that no one pays for and everyone demonizes you if you even try
◧◩◪◨
67. dayjah+cM4[view] [source] [discussion] 2025-05-08 03:33:46
>>pc86+4X2
You have superbly summarized the point I was trying to make regarding valuations.

Regarding your other point about pedigree: I’ve a high school level education. I was considered senior at this specific co. as I had played an important role in building the product (I’m a “classic case” of self taught generalist). I’m not at all clear on how I was selected to take part in that effort to be frank. It was fun.

◧◩◪◨
68. dayjah+CM4[view] [source] [discussion] 2025-05-08 03:39:04
>>foobar+qw3
For sure, as was the opportunity cost of redirecting our engineering teams to produce the competing product. That is, slowing the roadmap on our core product to deliver a tangential product.

Also, we had to factor into account user management, infosec bring-up, streamlining deployment, cross training teams, etc.

This was a large co., there was generally the belief that everything can be copied relatively quickly to reach something like parity. Obviously, the discovered path has tons more learning in it that the copier doesn’t benefit from when making subsequent decisions, which can result in lots of time lost.

◧◩◪◨⬒⬓⬔
69. pantul+ni5[view] [source] [discussion] 2025-05-08 10:26:26
>>therea+ri3
Thanks for answering! I also skipeed that the original codex now allows for other models (perhaps they pulled the open-codex part???). To be fair, I prefer Claude Code to both codexes.
◧◩
70. neets+Vw6[view] [source] [discussion] 2025-05-08 18:29:27
>>sagarp+TY1
It’s probably those big Fortune 500 corporate customers and getting a look inside their code bases or at least get to know their use cases
◧◩
71. shift8+Lc7[view] [source] [discussion] 2025-05-08 22:55:41
>>sagarp+TY1
I think it was foolish to buy it for so much money to be honest. I'm not sure how large its user base is, but that's more than likely something to do with it.

Remember that none of these tools can survive forever. Exiting was EXTREMELY smart of the founders here. Incredibly smart. I'd F off into the sunset now myself.

The reason is because they are built on top of VS Code and use Claude. So OpenAI can switch the LLM, cool, but at the end of the day there's no moat for the Cursors and Windsurfs of the world. OpenAI has the keys here because they have the proprietary LLM. This doesn't mean OpenAI will survive btw. I think Google will awkwardly win this race. They're so so so awkward though it'll take some time. It's painful to watch, but because they have the user base, they'll win. Hold that thought for a moment.

So new tools like Cursor and Windsurf will pop up all the time and do you think Microsoft will just sit by and watch? Nope. They'll update VS Code and Copilot and voila, the pendulum shift. As quickly as Cursor and Windsurf gained users, they'll lose them all again back to VS Code and Copilot. Copilot does indeed index your entire codebase - a rumor or misunderstanding by people. So as the dust settles, we'll see this change in usage. As LLMs trade places, we'll have a revolving door of fanbois and people arguing about what's better. What a rollercoaster.

What's really interesting here is that I think we're going to see a LOT of litigation. Back to Google. Look at what's happening with Google Chrome. Some genius thought they were in violation of anti-trust laws (well maybe they are). The reality is they are about to make it WAY WAY WORSE because Google Chrome was open source and should someone buy Chrome and decide they don't want to share...Well bye bye funding to virtual every other web browser. Congratulations brilliant legal system, you created a monopoly. It just wasn't Google's monopoly so I suppose that's alright.

So what I would bet is going to happen here is we're going to see OpenAI, Google, Microsoft, Anthropic, etc. all start to mess with one another here in a similar fashion. It's all going to be lobbying and legal battles. All over the place. It's going to make for an incredibly turbulent landscape. This is a very very expensive game.

THIS is why we're talking $3B. Someone is looking to protect something. It's not because of value. To think about it as a business value somewhere is wrong. Windsurf and Cursor aren't worth billions. Are you out of your mind? There's no justifiable way any rational human being would think that. When they're built on top of other tech and have absolutely no defensible position?? Heck no. It's not about their value individually, it's about their added value to these other companies. It's about bet placing. It's about protecting a larger business.

Take a moment to think what the world would look like if OpenAI must remain an open source business and or had to divest ChatGPT. They get Screwgoogled. Now ask why they're going to gobble up other businesses and why they keep raising all this money. I still think it's foolish, but it definitely seems like an existential threat that's lurking in there to me.

◧◩◪◨
72. sigmoi+EF7[view] [source] [discussion] 2025-05-09 05:11:47
>>thomas+2F2
They only get the preprocesses stuff that is sent to their api. But if you want to do complex coding tasks, you need the whole user interaction with the project and not just bits and pieces.
◧◩◪◨⬒
73. Mtinie+cT9[view] [source] [discussion] 2025-05-10 01:20:02
>>buffer+v22
Your comment only questioned why OpenAI cares about Windsurf from a first party perspective. I expanded the rationale for why the acquisition makes sense at a different level.
◧◩◪◨⬒⬓⬔
74. justan+l5a[view] [source] [discussion] 2025-05-10 04:30:16
>>indymi+UR3
what would someone with less or no experience do then? to be desirable?
replies(1): >>indymi+dKr
◧◩◪◨⬒⬓⬔⧯
75. indymi+dKr[view] [source] [discussion] 2025-05-17 00:44:52
>>justan+l5a
Making things fast with llms.
[go to top]