Everything bends to power, by definition. And laws can’t be impartial because they’re not based in hard science: terms like “murder”, “assault”, “theft”, etc. are ambiguous thus up to interpretation (e.g. is a scam theft? If so, what defines a scam? If “lying”, what’s the difference from “misleading”, or if there’s no difference, what defines “misleading”…)
My best idea for a solution is better education, so people don’t make bad laws then badly enforce them.
I don't mean to disagree with you in spirit, but profitability is pretty closely entwined with probability. So companies are chasing solving problems that more people have, even if it's for the wrong reason.
As the benefactor of an extremely rare disease, it's not exactly unfair when you look at it from a societal view. If you solve a higher probability problem, you are helping far more people.
The real tragedy isn't the allocation of the resources we have spare, it's that so many of our resources are not spare because billionares and corporations have hoarded it.
Without changing the percent of allocation, and only changing input resources by capturing it back from billionaires as taxes, we could be helping far more people including super rare diseases.
And if you take a step back and look at Covid spending, what it was spent on, and how much fraud was involved, it's absolutely maddening that the government isn't instead spending money on solving actual problems its constituents face. We basically just shoveled free money at anyone who claimed to have a business, to no real effect.
I don't know how much longer it will last but the US government invests significant resources into rare diseases in order to improve outcomes where the normal market wouldn't otherwise support the r&d.
A metric other than profitability seems like a terrible target for private research which (outside of a charity or cause-driven org) needs to justify its expenses.
In the US alone, we have dozens of grants, programs, and funding sources for things like orphan/rare diseases.
Those seem like pretty good ideas of the kind of targets we should have. But as you mention, those seem to only be considered for cause driven places or charities.
For a private company, QOL for someone doesn’t pay the researcher, laboratory chemical supplier, or lab’s landlord. Money does.
You won’t get a cure or treatment if you can’t discover it, get it approved, or distribute it, all of which costs (a lot) of money.
Inevitably, the bigger the quality of life improvement or more significant the impact = more money, so those measures are already indirectly considered.
It’s only incredibly rare diseases that aren’t receiving active research efforts.
The reason why it's less profitable is because it will help less people. If profitability didn't dictate what is researched, widespread diseases would get less researched and rare diseases - more researched, which would be a net negative.
IMO the issue isn't discovery and research, it's development. Unless companies foresee a good return for buying/licensing/etc rights to treatments, discovered drugs with potential just sit there.
What sucks is when drugs are deliberately not brought to market, but kept in portfolios, because it might impact sales of other existing cashcows. For example, Gilead has a history of staggering the release of new drugs only once their patents expire for similar drugs they already have on the market.
If you do that, you end up with all the problems that MCP attempts to solve: how to authorize using a standard mechanism, how to summarize operations in a better way than just dumping the OpenAPI spec on the LLM, providing structured input/output, providing answers to users that the LLM cannot see (for sensitive data or just avoiding polluting the context) and so on.
Authorization, in my opinion, is the big problem you need MCP for, though the current MCP Authorization spec still needs some refinement.
> It's like someone is claiming they unlocked ultimate productivity by washing dishes, in parallel with doing laundry, and cleaning their house.
In this case you have to take a leap of faith and assume that Claude or Codex will get each task done correctly enough that your house won't burn down.
"While LLMs are amazing, they can't run your business by themselves... We ground AI in tight guardrails and deterministic frameworks, optimizing LLMs to deliver enterprise-grade reliability. Trusted. Reliable. Secure."
this sounds like it's copy and pasted straight from an LLM
have you built stuff with LLMs before? genuine question because nondeterministic and deterministic workflows are leagues apart in what they can accomplish.
reply