Slava Kim

4.2K posts

Slava Kim

@imslavko

Katılım Şubat 2011

434 Takip Edilen3.3K Takipçiler

Slava Kim@imslavko·1d

@chongz @scale_AI @cartesia everyone knows the biggest red flag is attending the harker hs

English

713

Timothy Luong (Chongz)@chongz·1d

I didn’t stay anywhere longer than 2 years until my 5th job at @scale_AI, and the longest I've been anywhere is my current role at @cartesia. What’s counterintuitive is these "jumpers" are usually high potential (how they land the roles), but they have some traits that can hinder success in the wrong environments: - Some are talented but don’t know what they want. - Some want to be interested in their work but get bored quickly. - Some are overly interested in prestige. - Some should actually be founders. I think @HarryStebbings point is well founded, on average “jumpers” will end up more negative than not. But if you're able to find the few that end up being missionaries for your organization, they'll be some of the best contributors you hire. Some of the strongest emerging alumni mafias (e.g. @xai & @mercor_ai) almost entirely comprise of folks with tenure < 2 years, but you're going to miss some gems if you dismiss them out of hand.

Harry Stebbings@HarryStebbings

Single biggest red flag when hiring: Jumpers. Year here. Year there. When it happens 3 times or more for less than two years, I am out.

English

115

26K

Slava Kim@imslavko·2d

A new rule on our team: - can't request a review until CI tests are green - can't request a review until all Devin comments are resolved

English

226

Slava Kim@imslavko·27 Şub

@zeddotdev there seems to be some bug reloading the file that has been edited in zed, while also updated by claude code

English

290

Slava Kim@imslavko·25 Şub

@juiceropouch bro is about to get banned from entering south korea

English

ᴊᴜsᴜɴɢ@juiceropouch·25 Şub

first funemployment blog post! Reverse engineering the KakaoTalk app so I can build a Beeper Bridge jusung.dev/posts/kakao-ta…

English

825

Slava Kim@imslavko·23 Oca

@bryhoyt @itsandrewgao @cognition during the beta I had a similar moment where I asked them "Can you please make sure if the same file contains changes from 2 separate logical steps, the hunks are split" and they kinda said "it already works this way". Had to learn to trust it.

English

Bryan Hoyt@bryhoyt·22 Oca

@itsandrewgao @cognition awesome, exactly what I was hoping for. I might be unusual but if I'm anything to go by, I'd really recommend making that fact obvious in your docs / marketing if you haven't already. imo it's a big decision factor in how much to trust a review tool

English

149

Bryan Hoyt@bryhoyt·22 Oca

Having now tried this, it looks pretty great! The "group related changes" feature is huge. I have a question for @cognition: can I be sure that the grouped diff includes every single change somewhere? Is it deterministically complete? Or just an llm's idea of what might matter?

Cognition@cognition

Meet Devin Review: a reimagined interface for understanding complex PRs. Code review tools today don’t actually make it easier to read code. Devin Review builds your comprehension and helps you stop slop. Try without an account: devinreview.com More below 👇

English

3.7K

Slava Kim@imslavko·22 Oca

I was in the private beta for this product and the coolest part was seeing how they completely revamped the UI and the workflow UX based on the feedback in the beta group.

Cognition@cognition

English

633

Slava Kim@imslavko·31 Eki

The world we live in

English

578

Slava Kim@imslavko·8 Eki

buried in a thread but as a (backend) eng with interest in design and ux this is how I think about it my best past coworkers with the design sense were not gate-keepy either

Dylan Field@zoink

@mattaningram @DannPetty People SHOULD want to be designers. I've said it for years: design (and craft and brand and point of view) is how you win or lose as a business. This is more true now than ever before.

English

812

Slava Kim@imslavko·27 Eyl

Do you ever stop and think: I was born to orchestrate multi-channel conversational AI agents at enterprise scale

English

522

Slava Kim@imslavko·19 Eyl

@b_nnett @ajrgd @georgiarust3 1. the IPO filing gave an estimate for the future spend with the assumed optimistic growth 2. just storing and serving a bunch of user data in a big DB + S3 with back ups and redundancy costs $$$ 3. if you plan to train ML models (which they imply) it costs $$$

English

Bennett@b_nnett·19 Eyl

@imslavko @ajrgd @georgiarust3 So why is their compute cost in the millions?

English

Georgia Rust@georgiarust3·19 Eyl

we just upgraded Figma’s renderer to WebGPU ⚡️ WebGPU unlocks new rendering techniques and optimizations that will make Figma faster and enable us to ship richer visuals and smoother interactions. so proud of the team for making it happen! 👏 see how we did it 👉 figma.com/blog/figma-ren…

English

1.3K

125.8K

Slava Kim@imslavko·19 Eyl

Excited to see Figma roll out the WebGPU version. It was actually working internally for years, but my understanding is that older laptops and browsers were not compatible.

Georgia Rust@georgiarust3

English

1.2K

Slava Kim@imslavko·19 Eyl

@ajrgd @georgiarust3 This is not how it works, Figma renders everything in the browser. The only times it renders on the server is for stuff like thumbnails.

English

490

Alex Greenland@ajrgd·19 Eyl

I recently understood why your AWS bill is in the millions a month… you're rendering the canvas on the server side, and pushing pixels/gfx to the client. Which is great for supporting low-end devices and keeping the secret sauce — secret. But there must be a way to tune the compute and bandwidth.

English

4.1K

Slava Kim@imslavko·8 Eyl

$73M ARR before Windsurf, at $20M total burn

Cognition@cognition

Today, the baseline has changed drastically. While we are still in the earliest innings of AI code, agents are already doing real work alongside individual developers and within large enterprise engineering teams. As with many technologies throughout history, what first seemed a fringe theory quickly became an obvious reality. Over the past year, we have observed this shift, first with Devin and now with Windsurf. Before acquiring Windsurf, Cognition’s Devin ARR grew from $1M ARR in September 2024 to $73M ARR in June 2025, as usage increased exponentially. Our growth remained efficient throughout, with total net burn under $20M across the company's entire history.

English

186

39.2K

Slava Kim retweetledi

Cognition@cognition·8 Eyl

We’ve raised over $400M at a $10.2B post-money valuation to advance the frontier of AI coding agents. The round was led by Founders Fund with other existing investors including Lux, 8VC, Neo, Elad Gil, Definition Capital, and Swish VC all doubling down. We’re also joined by new investors including Bain Capital Ventures and D1 Capital. Two of our early investors, Christian Lawless of Conversion Capital and Emily Cohen of Neo, have even joined our team full-time.

English

144

174

2.2K

2.6M

Slava Kim@imslavko·4 Eyl

The internet infra has not accounted for the silksong seasonality.

English

597

Slava Kim@imslavko·4 Eyl

@ImSh4yy based

English