#Decagon vs Harvey vs Plaid(RO)

1 messages · Page 1 of 1 (latest)

iron abyss
clever yew
#

Harvey or Decagon fs

vale roost
#

it depends on how safe u feel ai is

#

i would choose here between harvey and plaid based on that bc they seem a better fit for u technically

toxic condor
#

since ur indexing for learning/team quality, harvey seems like best choice

lyric spade
#

Agent swe is not real swe work afaik

#

Its like a fde role

limpid dove
#

Even tho both companies could be dead in 2 years go with Harvey or Decagon

#

Prob decagon

ruby jolt
#

wait why not plaid

#

aren't these RSUs just fake cash

orchid crane
#

yeah probably consider potential of being pigeonholed in GTM / fdse if u take decagon

#

although fdse seems increasingly popular now and TC is pretty insane

limpid dove
limpid dove
orchid crane
#

but not if ur trying to do IC infra stuff in the future

limpid dove
iron abyss
limpid dove
#

As an fdse you’ll be vibe coding and shipping code fast

#

You’ll know how to quickly spin up short-term solutions for customers

iron abyss
#

oh sorry i thought you said fde is the way to go to be a good eng

#

i didnt see the idt

limpid dove
#

But you won’t actually be maintaining large systems or working on problems of scale

iron abyss
#

but yea i agree that's my only concern w decagon

hard niche
#

@north bobcat @young bloom GOATVEY

vale ivy
#

harvey contracts may end soon in which case company might struggle

limpid dove
neat sand
#

What can decagon do that ChatGPT can't do? What can windsurf do that ChatGPT can't do? What can glean do that ChatGPT can't do?

limpid dove
neat sand
#

Wdym

#

Windsurf had a good outcome for everyone involved 😂

#

It's pretty naive to critique all product layer ai startups as gpt wrappers, there's a nontrivial amount of engineering and product work required to get enterprise grade agents into production

#

Pretty sure decagon finetunes smaller models for specialized tasks

limpid dove
#

They didn’t sell for $3b

limpid dove
#

Ah cognition didn’t buy them for $2.5b

#

Cognition literally wouldn’t be able to afford them at that price

#

You’re getting confused by the Google licensing deal which didn’t even go to the employees

neat sand
#

Cognition paid the same valuation as google

#

Rumor is cognition paid 250m total at 2.5b valuation; the remaining employees had 10% equity

limpid dove
#

Also how does that even work? That would mean cognition only owns 10% of windsurf

#

So it was more of an acquihire if anything

neat sand
#

Yea it was an acquihire lol

#

Scott wu was pretty explicit in saying they wanted windsurfs gtm team

#

Either way all parties made out with all their equity vested at 2.5b

limpid dove
#

So ig you’re right

neat sand
#

Yea sounds like this guy got fucked bcuz he rejected the deepmind offer and then couldn't be made whole when cognition came in bcuz there was only so much in the deal + windsurf's remaining balance sheet set aside for the employee equity pool

#

At the time I remember reading a twitter thread that analyzed windsurf's cap table

#

Iirc google "bought out" equity of founders + vc + early employees and cognition swooped up the rest of the employee equity

#

Lol I think cognition realized this guy had 0 leverage since all the other founding eng alr left to deepmind and decided to fuck him

uneven roost
#

orz

alpine carbon
#

Did you have to do decagon in person?

jagged pond
limpid dove
#

Which values the company at half of what they were valued at in their last funding round

#

In other words, let’s say you were given $100K in RSUs that vest over 2 years in 2023. Then imagine working at the company for 2 years and being told you can only sell those RSUs for half of what they were worth, so $50k

iron abyss
#

the only upside to that is plaids valuation is a lot more realistic now so i’d be joining during the dip

jagged pond
#

Plaid was supposed to be acquired by Visa in 2020 for 5.3b but despite growing ~5-10x since then we're supposed to believe it's only worth 6.1b?

#

Pretty big upside there and the employees are well aware of this

limpid dove
jagged pond
#

Okay it was down 30%, that's still nontrivial

limpid dove
#

Antitrust?

jagged pond
#

Yeah antitrust

limpid dove
#

Fuck Biden

jagged pond
#

I'm a plaid glazer because they have a really good culture but not sure if i'd take it over my current role let alone decagon

#

Actually decagon is a sweatshop so maybe

limpid dove
jagged pond
#

I mean until very recently you could say the same for the stripe folks making bank in 2021/2022

#

Seems like tender offers are the new norm

#

For companies that don't want to IPO quickly

limpid dove
jagged pond
#

They do tenders twice a year with a 1 year waiting period

limpid dove
#

There’s always a risk in joining a private company

limpid dove
#

Also we’re talking about plaid

#

Op should def go to decagon or Harvey just for the clout

#

Since they’re already prev plaid

jagged pond
#

It just seems like this is what it's going towards

#

Because databricks is also doing yearly

#

I agree with building on your existing experience with diff firms

limpid dove
#

Maybe

jagged pond
#

Harvey is a question mark I've had a lot of friends leave recently

limpid dove
#

What happens when the VC money dries up tho

jagged pond
#

Decagon is straight hell but they prolly will raise

#

So good chance to make bank there

limpid dove
jagged pond
#

Once they hit ~10b they'll likely start giving some sort of liquidity to employees

#

That's what sierra is doing i think

limpid dove
jagged pond
#

Okay not yet but it is rumored

limpid dove
#

Which seems crazy bc decagon is certainly not a $10b company

jagged pond
#

I can't really say what they're raising but it's not 10b

#

Problem is that their direct competitor is

#

And another competitor raised at 1.8 with the same ARR

limpid dove
#

Don’t they technically compete with eleven labs?

jagged pond
#

Idk about eleven labs

limpid dove
#

They’ll be eaten up by oai anyway

#

Lowkey it’s wild to me that Anthropic is open ai’s biggest competitor

#

Yet they don’t even a multimodal model

misty apex
#

Riddle me this, if decagon is reliant on OpenAI for core functionality and OpenAI knows their exact prompts and is able to guess the surrounding architecture, how hard can it be for OpenAI to setup a direct competitor within 2 weeks

Genuinely curious not shitting on wrappers which I know have quite a lot of infrastructure Arnd to make it production ready but feel like they are just like iPhone apps which get wiped out when Apple makes it a widget

iron abyss
#

yea there is a lot of domain knowledge and business logic added on top of these api calls so it’s not as simple as a wrapper company

neat sand
#

Riddle me this 😹

stark marsh
#

as someone who also interviewed with harvey and decagon i would lean decagon mostly because of the better upside

#

but fwiw ive heard decagon is getting steamrolled by sierra rn hence their 996 culture

limpid dove
neat sand
#

Nah sierra is fine they will exit b4 oai eats their lunch 🔥

slender moon
vale roost
#

bro if the king of "i just want to coast" goes to decagon idk what the world has come to

#

u will be working 70 hr weeks 😭

#

dude go to plaid then

#

decagon culture is crazy lol

#

its stated clearly online that 6 days a week is the culture too lmfao

#

i dont think ull underperform at goog they dont expect 996 there

#

u didnt underperform

#

u got 3s thats not underperforming

#

give urself some credit dont self doubt ur way out of a job bro

#

yea i feel like u r psyching urself out

#

but i do think decagon is the wrong decision here for u

#

exactly none of them have a life outside their job

#

you will be spending like 75% of ur waking hours at work or going to work

#

thats fucking horrible