Sabitlenmiş Tweet
Krishang
225 posts
Krishang
@0ddmonger
granular thoughts and right of first refusal
Katılım Kasım 2022
62 Takip Edilen8 Takipçiler
had the misfortune to sit it on a vc meeting today for my friend’s startup. tier 1 indian vc
i can’t describe how retarded these guys are. also don’t understand what experience they have running companies to be advising founders on what to do. i looked them up most of them are mba grads that have worked all their life in vc
you’re a service provider of allocating your lp’s capital, can you do just that and shut the fuck up
no wonder the founders that know what they’re doing are raising directly from the source for mid 8 - 9 fig+ rounds
English
@thedankoe I think the winner in the next 3/5 years will the individual who can grasp and process information the fastest. Understanding with the implications in different verticals.
English
You as a single person have more power today than a 20 person company of the past. That's insane. The internet gave you the ability to learn anything. Social media gave you the leverage to reach anyone. AI is giving you the ability to create almost anything. Please don't waste it
English
I spent a couple of hours today trying to assess the feasibility of open-source models, running locally. At this point in time, it's extremely unrealistic to be running the Gemma 4 models on the higher parameter end. Plus fine-tuning it or deploying it is not easy task [for the long-run].
The reason why most consumer facing products are so easily plug and play is cause not a single soul will invest time into making it work [regardless of whether they're a free/paying user].
Now on the developer/tech-savy, forward-looking side of things, I'd say compute costs [A Mac mini or a super-optimized, mass-produced computer for AI] might reduce but running them will still need tinkering time and energy [regardless of costs]
English
@0ddmonger do you think most people will stick to free models / open source models?
and people that can afford it will get the best?
English
I’m wondering at what price do people stop happily paying for Claude or ChatGPT?
$500/mo
$1,000/mo
$5,000/mo
English
@thedankoe Every generation says that the next generation has gotten it easier. I beg to differ though. Yes it’s easier but there’s so many ancillary factors that’s made living, in general, harder.
That’s no reason to stop working though. Time to go all in!
English
Social interactions are honestly key to living in general. You need that night out with your gang. You need that random gym conversation about Drive to Survive. You need that conversation at the grocery store about food costs. You need that one sport you look forward to, meeting new people and enjoying the game together.
English
fascinating how many suggestions to this are solitary or even anti-social in nature
to me, a large part of mental fitness is social cognition - the ability to read the room, spark conversation, make new friends, flirt, etc
arguably as demanding on the brain as chess or a book
ted@tednotlasso
is there a mental equivalent of gyms but for brains? will this become a thing? do you think it will take form in writing clubs or a revival of Gilded Age social clubs, but all with strict no-phone policies (to avoid outsourcing any thinking)? something else?
English
@Codie_Sanchez I think about this a lot. Sure the first 5-10 days will feel good but what would you do after that? Watch shows? Go places [all needs a ton of cash], then what? Why not work at or on something that you enjoy/love.
English
@punk9059 Built a network graph for the Indian venture ecosystem. Over 9000 deals, across 5 years. Might help all you folks in fundraising and seeing who invests in whom. At a granular level.
dealverse-tau.vercel.app
English
@LeilaHormozi Wrote about something similar in my Social Constructs article. People seek for validation and those who are inherently people pleasers are so much more worse off in this case.
English
Too many people try to appease others before considering: Are these even people I want to be liked by?
English
the state of b2b saas in 2026: it's faster to vibe code your own tool than to onboard onto one that already exists
today i tried to set up a customer support saas for my startup — we're at the point where we need to reply to 100+ of client messages a day. the onboarding was so insufferable (3 hours of manual integration or hop on a demo call) that it was easier to just vibe code the whole thing than actually use their product
English
Hmm interesting, how long have you been using it for and what's the closure rate for issues?
The ticketing aspect of it is stored on supabase or just a local file?
If the agent has to run 24/7, don't you need a dedicated device for this as well.
But that's awesome that you got it done in like 25-30h of work.
English
vibecoded as part of my claude subscription, so it's ~$0
3h to code and 1 day to debug
ongoing token usage is negligible as part of minimax subscription $40/mo
looks better than paying even $200/month for 1 year
it's not as polished as intercom, but solves maybe 85% of my problems
English
@Toll3Y_ @Mericamemed The sad thing though is that Meta has shut shop on the Metaverse. Most would argue that the metaverse was driving the VR conversation on the consumer side. So who knows but 100% would love to try it out.
English
@0ddmonger @Mericamemed Yeah for sure, with the way those vr headsets/glasses and stuff are going I can definitely see it being a big thing but you'd probably have to dedicate a room to it, I don't think we'll get any nervegear type of things where you're stationary/brainwaves etc but this way 100%.
English
Let's run the numbers yea, How much did it cost for the vibe coded platform that you built, is it internal only? How long did it take for you to build it? and was the Opportunity cost of you building it vs focusing on your business worth it?
Would genuinely love to chat about it.
English
@0ddmonger wait, popular customer support saas tried to charge me 460/month for my volume + outcome-based pricing on the top
English
@Toll3Y_ @Mericamemed I cannot imagine how far ahead the Oasis seemed in 2017 with the state of Oculus back then. Definitely something that could maybe start off as a Gamer-hub arcade thing and then pass on to individuals.
English
@0ddmonger @Mericamemed First thing I thought too! Time for the Oasis! But for real, this would be amazing if a company did an entire collection of games for it, would be worth the expense.
English
@cyantist The hot takes getting old is something I cannot emphasize enough. Looks like brain rot has passed on from Instagram reels to Twitter. It's quite literally a 1000 tweets from 10k+ follower accounts tweeting about the same thing like parrots playing the telephone game.
English
The future isn’t written in training data.
Early-stage investing lives in what hasn’t happened yet. Humans are still the best at this and I can't imagine AI being better on any alarmist timeline.
Good luck training on that.
Not to mention, I can't imagine the best founders in the world wanting only to work with robots. This is a people business. Sure, some of the tedious shit we do will be automated, but we will, like all other industries, become more super human.
These "hot takes" are getting old and lack imagination.
GEOFF WOO@geoffreywoo
the venture capital bloodbath is coming and most vcs have zero idea agents will replace 90% of what associates and principals actually do: • deal sourcing through network analysis • due diligence via automated data mining • portfolio monitoring with real-time metrics • pattern matching across 10,000x more deals what exactly are you getting paid for when an agent can analyze every startup in your sector in 3 minutes? the entire industry is built on information asymmetry that ai just eliminated most funds will become algorithmic within 24 months the only vcs who survive are the ones who can actually build companies, not just write checks and send intros
English
I understand the innate psychological need for validation in terms of 1 and 3 (to an extent) but level 2 was and will always be something that will get better with time. Using claude/codex to augment your skills and be 20-30% better while taking the responsibility on yourself is the way forward.
The whole token spends flex is just bizarre to say the least.
English
Levels of AI psychosis:
Level one: You believe Claude's/ChatGPT's praise for you/your work is sincere and not just a ploy to keep you using the product
Level two: You measure your output in number of lines of code/number of github pushes (even though you did none of the work)
Level three: You stop caring about revenue/profit/growth and instead brag about money spent on AI tokens
Have seen level one and two for a while.
Seeing level three quite a bit more with founders posting screenshots of their Anthropic/OpenAI invoices to "brag" about how much money they are spending.
Bizarre world.
English
Codex has had some mixed reviews since the whole Claude 3rd party notice. If you prefer a conversational agent, then Codex is not your best foot forward due to its lack of character. It is extremely good at being short, execution heavy and to the point.
I haven't tried local models since I lack the infra for a good one [plus I wouldn't want to run one on my main pc]. Quite a few of them are still testing them out so using the extra usage while the X gang review local models is a decent strategy.
English
And yet there are zero AI teachers, zero AI doctors, zero AI lawyers.
Yafah Edelman@YafahEdelman
As far as I can tell diffusion is happening unusually rapidly in education, health, and law - which really puts a damper on arguments that regulation will slow AI diffusion.
English