320 karma · joined May 5, 2026
likely they focus on something orthogonal, like trust and security, permissions, etc
fwiw i work at one of their 'top customers' and found their ai to be really useful for spinning up basically ephemeral dashboards for anyone to use (aka, i dont need to file a request to the data team and wait for someone to get back to me or go back and forth). i just ask it to do the thing.
but profitability is a very narrow view on the labs, and investors are pricing in future free cash flow
you need to take into account the derivative, what is their growth? their opex and capex? their revenue growth? their moat? New product lines and applications?
Not saying that it's a homerun, but there is a case to be made for the labs being wildly profitable in 4-10 years
The price of a given point on the intelligence frontier is only going to get cheaper, likely 10-20x over the next few years.
software is easier to build -> the world builds more software
But yes, the act of translating spec into code as a 'programmer' is gone for good.
i add more value to my company because it lets me ship much faster, debug issues faster, onboard new people faster
total enterprise ai spend is climbing https://ramp.com/data/ai-index#adoption#overall
i know multiple stealth startups with 9 figure revenues and profitable and growing
lab arr is climbing
given the recent crackdown on token spend and from speaking to execs they are very focused on the exact ROI they will get from AI
Note that business revenues in competitive markets may actually not change that much, because if everyone is using the same models there's no competitive advantage from using AI. but if one of them didn't use AI they would fall hopelessly behind
This is because data annotation platforms have very low lock in and workers jump between them opportunistically, so they want to rely on as many (small) network effects as possible
Most data annotation comes from educated Western countries atp (the $2 meme is kind of old now since the industry has moved onto RL envs) where the annotators routinely earn multiple times the local median wage.
There has already been govt arms outsourcing the agent part to a vendor and it reduced hold times significantly without reducing quality (i.e. its possible)
But each govt arm can mess it up in their own way
- establish some sort of gateway that requires authentication connected to your personal information where they can rate limit or block you
- use ai of their own to handle the volume of requests (already happening)
why would they stop reacting when money is on the table and the technology to react exists?
the same robot that cleans your bathroom can also wash your dishes, fold your clothes, defend your home, fix your gutters, etc
(the bottleneck of which is the brain, aka software)
does the world use more or less software YoY? (LLMs counts as software btw.)
the differentiating factors have changed but the underlying demand has only increased.
doesn't feel like it, collaboration is more important than ever for high level direction even if nobody pair programs anymore. relationships feel more important than ever especially customer ones. theres no RL environment for competing in business
> We consume more food than ever, but the % of people who work to produce food is far smaller than it was 150 years ago.
Yes, because there is a limit to human desire for food, as we tend to get full. is there a limit to humanity's appetite for software? I would posit probably not, because software is infinitely abstractable and can be applied to any industry and domain
the price of code is falling but the value of differentiated ideas and market alpha is rising. human attention is scarce - as the mini-ceo with your army of agents, how will you compete with the other mini-ceos with their army of agents that can produce software at a whim?
"They" figured out that it's a lot easier to hillclimb AI on verifiable digital tasks, and it just so happened that white collar jobs tend to pay more
Ai will decimate customer support jobs but those are not considered highly paid
If robotics continue to progress, we'll have all the robots to do our dishes and clean our bathrooms as you wish
ai has produced some of the fastest growing companies of All Time and plotting the trend leads to very profitable businesses in the future (which is the whole point of investing)
i think it would be a more nuanced and interesting conversation to talk about 1) how growth will change and 2) how the rapidly falling cost of intelligence will affect the different players in the ecosystem
(not a bullpost for the labs, i think their situation is precarious)
the new era of swe will demand more and more long term business + product direction and relationship building skills
right now the meme is we all became managers of agents. the trajectory leads us to all becoming mini-ceos directing our company of agents
Software allows us to build essentially infinite abstractions with infinite room to run. cheaper and better models released to the world means higher competition which drives up the cost for talent
however i think the value of 'hand crafted code' drops to zero unless we convince consumers to purchase software as a status symbol like rolls royces (seems unlikely, unless there is a concentrated marketing effort)
Tbh i dont see software demand being saturated anytime soon (which is why im bullish SWEs, and we're seeing both demand and compensation increase as of late), we keep moving up abstractions and increasing ambition and scope for new projects.
If all the companies use {frontier model} then your competitive advantage is what you have that is not legible to the model
consumer subscriptions are subsidized but the companies make very healthy margins via api pricing, which is what enterprises use, and cost per intelligence is dropping rapidly
they cannot be held accountable - therefore, you need the skill and expertise to verify outputs, or set up a verifier
they have a limit to their context window - therefore, you need the skill of knowing what context is useful and when
they cost money and time to run - therefore, you need the skill of learning what parts can be handed off to code, when you can use a cheaper model, when you tradeoff accuracy vs latency
they are nondeterministic - therefore, you need the skill to set up guardrails, knowing when to reach for the llm vs. code depending on the problem
they are not omniscient - therefore, you need the skill of setting up the processes and infra to supply legible context to the llms
there are many different models - therefore, you need the skill of understanding model tradeoffs and building proprietary evals
etc etc
they cannot be held accountable - therefore, you need the skill and expertise to verify outputs, or set up a verifier
they have a limit to their context window - therefore, you need the skill of knowing what context is useful and exposing that
they cost money and time to run - therefore, you need the skill of learning what parts can be handed off to code, when you can use a cheaper model, when you tradeoff accuracy vs latency
they are nondeterministic - therefore, you need the skill to set up evals, guardrails, knowing when to reach for the llm vs. code depending on the problem
they are not omniscient - therefore, you need the skill of setting up the processes and infra to supply legible context to the llms
etc etc
its clear they own the mindshare around this type of primitive which is a massive premium
seems like they are aiming to provide both inference and RLaaS for american companies and western govts. even if they never fully beat deepseek if they get close enough the fact that they're American will help them close deals