Vibe @realfastai scrolling through lunch options. tinkering.
Flibbertigibbeting
Joined January 2021
- Tweets710
- Following926
- Followers294
- Likes2.2K
I’ve always imagined the roads in a dystopian future to be good, for some reason. Yes there would be weird mobs and traffic and hitchhikers, but never thought of the potholes.
Sage retweeted
if you experience reality in this way you should not.
the world is situationally and cyclically a dumpster fire. always has been always will be. kids will inherit the parents experience of it. implicitly and explicitly.
the deciding factor imo is not the world, it's the nature of the parents relationship with the world.
Sage retweeted
I was on Zerodha’s Subtext recently, talking to @pranavmanie and @bebhuvan about what an AI-native firm is, and the part people kept DMing me about afterwards was the few minutes on the middle tier of talent disappearing. This essay is the longer answer.
sidu.in/essays/after-ai-ther…
Sage retweeted
If you're a CXO putting serious money behind AI, which of these is bothering you already?
Sage retweeted
the one line answer to why AI rollouts fail - the KPIs tied to the rollout aren't oriented on getting work done
Sage retweeted
first million $ of revenue from US geo done @realfastai
we are now north of 7 figures of rev each from 2 geos, with the rest at 5 and 6 figures
Sage retweeted
Another reason I’ve been unphased by VC pitch weirdness is: high performance looks weird
And it looks weird in founders too—you want investors on your side who understand that
The world of mediocre performance is designed to create a sense of order, surface level politeness, predictability, warmth
When you push to the extremes of performance, and operate from first principles, the outcomes look alien to many people
It takes weird people to operate at this level
The last person you want on your board is a conformist bureaucrat who doesn’t understand the extreme chaos of running a high growth startup—because the reality will scare them and then you’ll have to put on a performance to manage them—distracting you from getting any real deliberation done
This dynamic is not super obvious because very smart, very weird people have a lifetime of practicing seeming normal on podcasts, in public, etc.
But in the environments where these people get together—it’s a totally different wavelength of communication, chaos tolerance, contrarianism, intensity
As a founder you also need to be able to identify weird and spikey people in hiring
It’s super important to understand how to work with these people and to understand why they think the way they do
It usually does not come from a bad place—it comes from consequentialist ethics and extreme optimization
Overheard a recruitment conversation @realfastai
“So, how were his jsonLs?”
“Terrible, no steering”
Welcome to the future, ig.
Yeah this makes too much sense. In the world of consulting, slide decks are so 2024. Every client needs a live html doc with dashboards / status updates that they can scroll through and make decisions.
This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your browser. I've also had some success asking the LLM to present its output as slideshows, etc.
More generally, imo audio is the human-preferred input to AIs but vision (images/animations/video) is the preferred output from them. Around a ~third of our brains are a massively parallel processor dedicated to vision, it is the 10-lane superhighway of information into brain. As AI improves, I think we'll see a progression that takes advantage:
1) raw text (hard/effortful to read)
2) markdown (bold, italic, headings, tables, a bit easier on the eyes) <-- current default
3) HTML (still procedural with underlying code, but a lot more flexibility on the graphics, layout, even interactivity) <-- early but forming new good default
...4,5,6,...
n) interactive neural videos/simulations
Imo the extrapolation (though the technology doesn't exist just yet) ends in some kind of interactive videos generated directly by a diffusion neural net. Many open questions as to how exact/procedural "Software 1.0" artifacts (e.g. interactive simulations) may be woven together with neural artifacts (diffusion grids), but generally something in the direction of the recently viral xitter.cf/zan2434/status/2046982…
There are also improvements necessary and pending at the input. Audio nor text nor video alone are not enough, e.g. I feel a need to point/gesture to things on the screen, similar to all the things you would do with a person physically next to you and your computer screen.
TLDR The input/output mind meld between humans and AIs is ongoing and there is a lot of work to do and significant progress to be made, way before jumping all the way into neuralink-esque BCIs and all that. For what's worth exploring at the current stage, hot tip try ask for HTML.
Sage retweeted
Been living this. Built toss for exactly this loop — self-deploy on Cloudflare or Vercel, get an expiring/permanent share link for every HTML artifact your agent generates.
github.com/piyush0609/toss
Sage retweeted
Most managers earned their role by being one of the best ICs on the team.
It's time to be one again.
Most Managers come up the ranks as top ICs.
So in a pre-AI world, when someone said "this is going to take 3 weeks," they knew if that was true because they had done it themselves.
Now think about post-AI work world.
Most current managers have never been an IC with AI at their disposal.
They don't really know what's possible or how long stuff should take.
We see this play out at Segwise all the time.
A founder or head of growth finds us, the value prop clicks.
AI creative tagging, agentic creative analytics, competitor insights and creative generation.
All makes sense. So they send us to their team.
And then they step back.
Then the team evaluates.
But look, most people don't want change.
Their work life is fine right now. They have a process for tagging creatives in spreadsheets, they know how to pull reports from each ad network manually.
It works fine.
A new tool means relearning how they do their job differently.
It could be better, but it is definitely going to be different.
The best teams look forward to such exploration.
But there are also teams which do not want this change, and they go back and say - "Hey I don't think we need this."
In the meantime, our CS team has been chasing them to get on training calls and trying to show them how to get the best out of the free trial.
Every AI founder has dealt with this.
The manager takes that at face value because they don't have the instincts to challenge it anymore.
They haven't used the tool.
They can't tell the difference between "this genuinely doesn't work" and "I just don't want things to change."
Because they don't have the IC instincts for the AI version of the job they were great at.
Change creates chaos and most people don't want chaos in their work life.
But if you're a leader evaluating AI tools right now, you have to become the IC again.
You were one of the best with dumber tools.
You have to be one again, at least until your instincts become sharp again.
Sit with the new AI tools. Do the actual IC work.
Get a feel for what's possible now, not what someone told you is possible.
Otherwise you'll keep hearing "it's not ready" while your competitors bolt ahead.
Sage retweeted
A @realfast Agentforce engagement was just covered by @salesforce as a customer story!
We were able to dramatically improve metrics for our customer, TASC, using Agentforce.
(link below)
Sage retweeted
🚀🚀🚀 We're building real-time cross-border payments at @pay_glomo one of fintech's hardest distributed systems challenges.
Hiring Product Engineers (5-12 YOE) in Bengaluru:
Tech stack: Ruby on Rails, React, PostgreSQL, Redis What we need: Payments/fintech experience, ownership mindset, 0-1 startup energy
DM me or Apply → lnkd.in/gvhEMb7H
#FinTech #RubyOnRails #BengaluruJobs #Hiring
Sage retweeted
MY MIND HAS JUST BEEN BOGGLED.
I just tried the early release of OpenAI's new DeepResearch feature (rolling out later today to the $200/month Pro users).
I've been working in the CRM software industry for 30+ years.
It's not just that I've had court-side seats to the game, I've been on the court, doing my best to play the game. First in vertical CRM (my first startup) and now as co-founder of HubSpot. I've had some modest success and I feel like I have a pretty good handle on things in the industry.
That's why OpenAI's new DeepResearch feature boggled my mind.
I asked it create a detailed research report including competitive analysis, positioning, growth, product strategy and AI vision for the industry.
What it produced was an 11,000 word report. With data. And citations. And tables. And genuinely great insights -- including some I hadn't really thought of before.
What has me excited is not just that it can produce this kind of output (though that's pretty cool). What has me excited is that we'll be able to use this kind of output as *input* to a subsequent step in an agentic workflow.
Because the future is about agent composability. Being able to pull together pieces and put them together into a larger whole. The same way we build teams to work together to tackle higher order missions and goals.
This has been what I've been dreaming about for years. It's finally starting to happen.
We are seeing more and more of the future -- and it's happening quickly.
DISCLOSURE: I'm a small investor in OpenAI, and also a big fan.
Sage retweeted
We had a Mckinsey co-founders retreat in Goa this weekend - fondly called the Mckinsey Mafia
Hands down the best professional network I've had the privilige of being a part of!
Incredible to meet an amazing set of founders, at all stages and walks of life - From early ones like us to scaled giants like @naiyyasaggi, Rizwan (of Citius), @dhavalshah24, @sidtreebo and many more
Love the fact that this all happened organically - 🙏many thanks to all the founders who came forward to make this happen
P.S. I cleaned house at poker! 🥳
Sage retweeted
RIP Social Apps, 2005-2023
As of iOS 18, friend-based contact sync apps are basically dead on arrival. The number of people consenting to the new permission has nose-dived—so there is no way to get density in a meaningful way.
But as one door closes, another opens up:
From the handful of dashboards I've seen, the retention rates on AI companion/girlfriend apps are extraordinary.
Instinctually, I thought the entire space was bullshit. But it's now clear that the ability to craft relevant, empathetic responses to any user input is a powerful lever to activating users and entrenching them into a relationship, albeit artificial.
For better or for worse, LLMs might be the new Contact Sync. If we can't find our friends on apps anymore, people will find something else to talk to.