Chris Risebrough

@ethanzhou @LandseerEnga big cycles guy

its officially been one month since i started working in SF, here are a few things ive noticed: 1. basically everyone here is in tech lol. 2. "what are you working on" has replaced "what do you do" 3. the weather. 4. the lingo here is so funny, my boss doesn't ask if i have time, he asks if i have "cycles" 😭 5. everyone's either raising, just raised, or about to raise 6. ai billboards (and planes?) 7. pretty friendly people; was on caltrain and some random dude clocked that i worked in tech and asked for my linkedin😂 one month down and genuinely loving it so far. if you're in the city let's connect, always looking to meet more people!

@farazdotai claude hallucinating a speed up gain because it switched a 6 and an 8

Witnessed two interns go through a series of graphs, analyze the ops and shapes of tensors like by line Admirable they take the time to learn, but claude could have made a report and explain in 0.001 of the time

@hxyden finding out some anon IDs get overwritten every time an app launches smh

Your app does a thousand things under the hood you never got to see. Until now. Every network call. Every frame drop. Every byte of SQLite touched and mutated. You, your team, your agents. Nobody flies blind anymore.

@tryrevyl blood sweat and tears put into this but at least they weren't mine

On top of it

@georgecurtiss revyl x helix local phone db written in scratch coming soon

@ChrisRisebrough Is this the new meta?

all databases will be rewritten in Rust not us though, we're not retarded so we started with it

@hxyden muy bien

Mapped Duolingo's full onboarding in Atlas. 16 steps, 70 screens, 26 paths, end to end. Goldmine for anyone trying to understand what a masterclass in onboarding looks like

@zachtheoldham @cognition how many ads do I need to buy for a gravity hoodie?

@cognition how much do we need to spend/month on Devin to get a hoodie

@HaroutKhack they didn't deserve you anyways

> Interviewer asks a hard competitive programming problem > I miraculously solve it > still didn’t pass the interview

@dhruvr_43 Heard waterlooworks needs some devtooling

@lucasdontz how I be joining the google meet with the random account manager who's trying to get me to increase our usage

Status games. What makes dialogue realistic. How sunglasses avoid the submission of the eyes. No time to think about fulfillment, just go out and fulfill stuff. thoughtservice.substack.com/p/status-and-t…

@sfspacesquid company smart cars could be the move

Midnight purple Mondays

@michaelfromyeg Playoff time brings the best out of people

y'all know I’m Canada down

@anamhira let the equity vest first at least

You know its a good sales call when the company asks to acquire you.

@hxyden naming orgs like a junior ib analyst naming powerpoint s

First screen after signup. Before and after. Value before paywall > paywall before value

@zachtheoldham @leojamartinez zach ted talk when

One of my biggest motivations when I started building an ad platform with @leojamartinez was to create profit-focused campaign objectives (and other cohort, LTV, new customer and incremental objectives). Many of the best performance and affiliate marketers are here on X so this is far from a novel concept for them, yet it still baffles me that such a large % of the digital advertising spend existing in the world today is strictly “ROI” focused with no consideration of profit (specifically contribution profit : Revenue - COGs - Shipping/Handling - Payment Processing - Taxes - any other variables). Thanks to the likes of @AY_Orbach and @aryehMaxx from @triplewhale or @TaylorHoliday from @CommnThreadCo (along with dozens of others) who brought education and tooling to the community, this concept is fairly well understood now. And it is far from being the best or even most useful way to measure marketing success. But despite the awareness among the best marketers around, an insanely large group of marketers still consider ROI as a North Star. At a brand level this may make sense, but when applied to different campaigns, selling different products to different groups of people, this is completely flawed. A simple example of this is as follows: Campaign 1 is focused on selling socks with a 50% gross margin (meaning a 2x return is needed to break even after all variable costs are accounted for) Campaign 2 is focused on selling hoodies with a 20% gross margin (a 5x return is needed to break even) Many people still look at ROI in their ad account, and optimize based on that metric. Which makes no sense whatsoever. Let’s say Campaign 1 has a 3x ROAS, and Campaign 2 has a 4x ROAS. In the above example, Campaign 2 is losing money, even though it looks like it’s performing better. A media buyer then shuts off or turns down budget on the first campaign (even though it’s the only one that is profitable), and reallocates budget into Campaign 2… killing the brands overall profitability… The reality is that this is a very crude example, and in order to be most accurate, you need to be able to understand the profitability of a cohort, and apply it at the ad, ad group, and campaign level to optimize accordingly. And while there are technically ways to do this on @Meta or other platforms by sending custom data back to the platform and using these custom datasets as objectives, the lack of this being a default option, is a huge missed opportunity. In fairness to Meta and others, they have incredible people working very hard on delivering better performance for advertisers, so I hope that these ad engines are eventually trained on these more “rational” objectives, but for now the industry will have to continue educating each other on these more advanced strategies for maximizing performance from the ad gods. And as we scale up Gravity, I promise that we will dedicate data, ML and engineering resources toward pioneering default advertising objectives that are most aligned with your bottom line. P.S. I know that there are many other ways that people think about and consider performance and measurement, and how wildly it differs across individual companies and industries, and lastly, that this example is uncomfortably primitive. But, I hope that it communicates the point that platforms can make it a lot easier for advertisers to buy the types of results they want than the current paradigm

join the best, most cracked canadian team in sf <>

@LandseerEnga HR case pending

Models are smart enough now. They don't need complex frameworks - just simple tools: screenshot, input, swipe. Watch claude code navigate instagram and leave a comment on my latest post.

@LandseerEnga @sfspacesquid race against sunset

Have you thought to yourself: “I wish I could go on a group ride with @sfspacesquid to hawk hill” Today is your day - join us on the south side of Golden Gate Bridge at 6:15pm! Dm me