Malav (@malav399) header
Malav (@malav399) profile picture

Building consumer agentic factory | RL environments | Simulation | Human belief | Consciousness

Bangalore
Joined April 2016
Photo by Malav (@malav399)Photo by Malav (@malav399)Photo by Malav (@malav399)Photo by Malav (@malav399)Photo by Malav (@malav399)Photo by Malav (@malav399)Photo by Malav (@malav399)Photo by Malav (@malav399)Photo by Malav (@malav399)Photo by Malav (@malav399)
We will see 2x growth in venture studio than we did in last 3 years altogether. Scaling single-consumer app will be scary. Venture studio with proprietary intelligence will beat at pace, operating cost eventually. Also revenue per token with shared intelligence will be atleast 2x higher than the prior. Still lot of moving parts in studio model but eventually creating disproportionate outcome. (I also mean distribution here - not only development)
1
120
Malav retweeted
So many times I have had this interaction with normies where they don't understand what the jackfriks I'm working on as I'm just making random things for fun and messing around with things And to them whatever I'm doing looks like a complete and utter waste of time, it has no function, no direct rewards, it's not making money, it's not successful But so many times those projects evolve or lead to things where there is a massive payout or success after But most people don't have the ability to see that far ahead, and they're risk averse so even if you tell them "I don't know maybe this becomes something, or maybe not, it doesn't really matter" that's simply too uncertain enough for them to accept. Which is of course why they love employment so much, it's stable and certain. The problem with normies is that they will quickly shut this creative play down in themselves and even worse in others. And that way whatever that thing is, it won't ever grow into anything because they killed it before it ever got a chance, proving their point Almost all my projects started by just messing around with no real goal, just playing for fun But it requires some kind of peacefulness because the moment you start putting a requirement of success on to something new, the thing instantly dies Creativity is really magical, and it messes with you if you try control it or expect things from it
A lot of the most successful people I know spent years of their life just bumming around, before they suddenly exploded into action and did something great. I don't think that's a coincidence. noahpinion.blog/p/bumbition
241
204
65
3,508
199,572
We really underestimate the laziness of humans AND buying the convenience, barring accessibility and skills. Canva was never out of scope or too complex yet there are so many agencies. Models will get better eventually solving at and for the upstream of the vertical. But most value will be created by ones that ace last mile delivery. Consumer is barely scratched!
1
63
Ever since absurdity got slapped with too many options, creativity has barely been able to find the light. Absurdity makes you think in extremes, which comes at the cost of partial failure. As a result, we’ve subconsciously developed a template bias. We have an idea of what the best should look like, and we’ve started calling it taste, when it’s often just references in disguise. Unfortunately, this is true across the board, from media, technology, design and culture. We've started treating the accumulated evidence of what worked yesterday as the boundaries of what can work tomorrow. Creativity is a function of extreme contraction thinking, or absurdity, and working backwards from there.
1
91
AI-native service firms are huge, but I don't think we talk enough about AI-native consumer firms. Bending Spoon has proven how operational efficiency can maximise scale and margins. As development gets cheaper, scale & margins become a function of proprietary intelligence, model harness, revenue per token, and optimising per-unit attention (LTV). Most value will be created at the end of the value chain by optimising models for the best user experience - just the way Higgsfield did.
1
184
Introducing Augmentic: Our first step towards photoreal robot simulations. Trained to take G-buffers from simulators like Isaac Sim and render a photorealistic frames. First results below 🧵
13
11
5
64
7,961
Criteria for right to win is being rewritten now. It’s not only inherent experience will win big but ones who can run 10,000 experiments at scale to redefine the category and how it is approached. Signals shall always be the journey to insight and ofc outcome and not only labels. The way an individual connects the dots and produces signal relative to resource will have greater chance of winning than an individual with work experience from the same industry. Next Disney or Pfizer will be built by someone who has never worked there!
105
Data infrastructure for bio alone will be worth trillions.
1
173
Post labour economics of white collar aka tier 1 audience will unleash the mental health category like insane. We are headed to unimaginable identity crisis.
171
Indian manufacturer's data is so difficult to source! They have 25 years of data - all on paper. Still works that way. Pretty much >80% of manufacturers in Gujarat, at least. What is the best way to digitize these? 🫠
3
213
Next big thing for frontier model and physical AI will be learning human behaviour at neurological level. User / population simulation cannot truly predict the subjective factors of decision making. The objective lens of simulation may not be completely accurate. Especially true for consumer - no user, for any consumer application- commerce, media, finance, health etc. decides basis data but emotion first. Otherwise, mostly true, for enterprise. Current approach like @simile_ai’s makes sense for certain verticals like healthcare. Forecasting human behaviour cannot be objective.
2
173
Intelligence getting cheaper & available in thin air doesn’t truly imply better intent or use. History has proven- more information has led to decline in common sense rather otherwise. Per capita stupidity will be at peak Signs you can see in comments section of reels 🫣
1
1
150
Once they start using agents it’s going to be terrifying brain rot/Psyop wedge to prevent it. It won’t be limited to Halku paglu - AI video 😭
1
43
So how many personal agents will live through the first-scroll on WhatsApp/iMessage effectively? And human chats later Open router/superhuman of personal agents? Are we really going to address it as personal agent? 🙃
1
97
Micro-apps say like speaking practice, lookmaxxing or Cal AI type - can be assistant-first WhatsApp/iMessage interface. It spike up engagement and retention by atleast 2x.
2
1
139
It's over for Clippers 🤯 Opus 5.5 - One-shot clip and edit 🔥 youtu.be/a00xIn5kwhM?si=b70T… Asked to clip 1 min from the link and edit it like the reference video provided.
1
1
3
286
Astra’s 3D rendering ability + opus 5.5’s + H3 max should really introduce a new format of interactive entertainment/media. Will be sad if it doesn’t happen. Short on bandwidth else would have tried it out
2
164
Agent harness for filmmaking, that too, vernacular first will be massive. UX cannot be like other video generation platforms - should be loveable like. Subjectivity and treatment of frames can eventually be harnessed with more data.
4
160
- Niche-sub-culture, - Long format content - Curated (to extent- personalised) feed/internet - Characters IP: multi format (like minion, hello kitty, Mario) - Interactive media IPs - ofc, analog experiences We will see significant deal flow in entertainment sector in 2027.
102
If you consciously audit your current or recently acquired beliefs - You’ll realise the line between our consciousness and our algorithm curated feed have gotten blurred. Beliefs beyond feed is rare. The evolution of our beliefs has mostly been induced by alot more amplifier nodes than we can think of. Volume + audio has the highest impact of all. Learning about evolving human beliefs is key for AI safety and RL environments for agents to go beyond human skills, objectively.
85