985 karma · joined September 15, 2010
An LLM which makes a tool call to a function called `ride_bike`, where that function is a different sort of model with a different set of feedback mechanisms than those available to the LLM, is NOT the same thing at all. The LLM hasn't "learned" to ride the bike. The best you can say is that the LLM has learned that the bike can be ridden, and that it has a way of asking some other entity to ride on its behalf.
Now, could you develop such a model and make it available to an LLM? Sure, probably. But that's not an LLM. Moreover, it involves you, a human, making novel inroads on a different sort of AI/robotics problem. It simply is not possible to accomplish with an LLM.
I mean, services _could_ make it harder to use LLMs to interact with them, but if agents are popular enough they might see customers start to revolt over it.
Using agents that interact with APIs represents people being able to own their user experience more. Why not craft a frontend that behaves exactly the the way YOU want it to, tailor made for YOUR work, abstracting the set of products you are using and focusing only on the actual relevant bits of the work you are doing? Maybe a downside might be that there is more explicit metering of use in these products instead of the per-user licensing that is common today. But the upside is there is so much less scope for engagement-hacking, dark patterns, useless upselling, and so on.
You're asking someone to answer this question in a text forum. This is not quite the gotcha you think it is.
The distinction between "knowing" and "putting into language" is a rich source of epistemological debate going back to Plato and is still widely regarded to represent a particularly difficult philosophical conundrum. I don't see how you can make this claim with so much certainty.
I happen to think there is validity to the statement if you control for other actuarial factors. But if you don't think that makes sense as a lens through which to look at the problem, I won't quibble, even though I disagree. We're also only talking about drug prices here, which is a small portion of overall healthcare spending.
In any case, the central point, that insurers benefit from higher prices, still stands.
Yes. As I wrote above, insurers compete on premiums, and they do do so by using rebates to subsidize those premiums by spreading patients' deductibles across the insured population. As far as profits go, I can't speak to regulatory issues since they will vary by state, but in any case the same critique would apply if insurers are pocketing a fixed percentage of a larger amount.
Re your second point, it completely twists my point and is largely irrelevant. Yes, older people paying the same premiums as younger people is a counter-argument in that older people are more likely to need healthcare, but the central point is that people who have to USE their insurance (i.e. sick people) subsidize the premiums of people who don't (healthy people), and this critique applies regardless of age. Now, one could argue that the structural factors that control costs across age cohorts counterbalances this phenomenon. And I'd agree with you! But that doesn't negate the original point that insurance companies benefit from, and advocate for, high sticker prices.
They can use the rebates they get from the providers to subsidize the insured, allowing them to offer lower premiums and gain market share. This is what people mean when they say "In America, the sick people pay to subsidize the health care of the healthy people".
Of course, that above only applies if there is competitive pressure. If there is no competitive pressure (e.g. in states with only one or two insurers), they can keep premiums high and book as profit the difference between what the patient paid out and what the patient would have paid out in a lower-cost no-rebate world.
For data MCPs, we use remote MCPs that are served over an stdio bridge. So our configuration is just mcp-proxy[0] pointed at a fixed URL we control. The server has an /mcp endpoint that provides tools and that endpoint is hit whenever the desktop LLM starts up. So adding/removing/altering tools is simply a matter of changing that service and redeploying that API. (Note: There are sometimes complications, e.g. if I change an endpoint that used to return data directly, but now it writes a file to cloud storage and returns a URL (because the result is to large, i.e. to work around the aforementioned broken factor of MCP) we have to sync with our IT team to deploy a configuration change to everyone's machine.)
I have seen nicer implementations that use a full MCP gateway that does another proxy step to the upstream MCP servers, which I haven't used myself (though I want to). The added benefit is that you can log/track which MCPs your users are using most often and how they are doing, and you can abstract away a lot of the details of auth, monitor for security issues, etc. One of the projects I've looked at in that space is Mint MCP, but I haven't used it myself.
> What's your hit rate on researchers actually converting LLM explorations into permanent artifacts vs just using it as a one-off?
Low. Which in our case is ideal, since most research ideas can be quickly discarded and save us a ton of time and money that would otherwise be spent running doomed lab experiments, etc. As you get later in the drug discovery pipeline you have a larger team built around the program, and then the artifacts are more helpful. There still isn't much of a norm in the biotech industry of having an engineering team support an advanced drug program (a mistake, IMO) so these artifacts go a long way given these teams don't have dedicated resources.
> Do you think this pattern (LLM exploration > traditional tools) generalizes outside domains with high uncertainty?
I don't know for sure, as I don't live in that world. My instinct is: I wouldn't necessarily roll something like this out to external customers if you have a well-defined product. (IMO there just isn't that much of a market for uncertain outputs of such products, which is why all of the SaaS companies that have launched their integrated AI tools haven't seen much success with them.) But even within a domain like that, it can be useful to e.g. your customer support team, your engineers, etc. For example, one of the ideas on my "cool projects" list is an SRE toolkit that can query across K8s, Loki/Prometheus, your cloud provider, your git provider and help quickly diagnose production issues. I imagine the result of such an exploration would almost always be a new dashboard/alert/etc.
[0] https://github.com/sparfenyuk/mcp-proxy - don't know much about this repo, but it was our starting point
Anyway: a lot of earlier stages of drug discovery involve pulling in lots of public datasets, scouring scientific literature for information related to a molecule, a protein, a disease, etc. You join that with your own data and laboratory capabilities and commercial strategy in order to spot opportunities for new drugs that you could maybe, one day, take into the clinic. This is traditionally an extremely time consuming and bias prone activity, and whole startups have gone up around trying to make it easier.
A lot of the public datasets have MCPs someone has put together around someone's REST API. (For example, a while ago Anthropic released "Claude for Life Sciences" which was just a collection of MCPs they had developed over some popular public resources like PubMed).
For those datasets that don't have open source MCPs, and for our proprietary datasets, we stand up our own MCPs which function as gateways for e.g. running SQL queries or Spark jobs against those datasets. We also include MCPs for writing and running Python scripts using popular bioinformatics libraries, etc. We bundle them with `mcpb` so they can be made into a fully configured one-click installer you can load into desktop LLM clients like Claude Desktop or LibreChat. Then our IT team can provision these fully configured tools for everyone in our organization using MDM tools like Jamf.
We manage the underlying data with classical data engineering patterns, ETL jobs, data definition catalogs, etc, and give MCP-enabled tools to our researchers as front end concierge type tools. And once they find something they like, we also have MCPs which can help transform those queries into new views, ETL scripts, etc and serve them using our non-LLM infra, or save tables, protein renderings, graphs, etc and upload them into docs or spreadsheets to be shared with their peers. Part of the reason we have set it up this way is to work through the limitations of MCPs (e.g. all responses have to go through the context window, so you can't pass large files around or trust that it's not mangling the responses). But also we do this so as to end up with repeatable/predictable data assets instead of LLM-only workflows. After the exploration is done, the idea is you use the artifact, not the LLM, to intact with it (though of course you can interact with the artifact in an LLM-assisted workflow as you iterate once again in developing a yet another derivative artifact).
Some of why this works for us is perhaps unique to the research context where the process of deciding what to do and evaluating what has already been done is a big part of daily work. But I also think there are opportunities in other areas, e.g. SRE workflows pulling logs from Kubernetes pods and comparing to Grafana metrics, saving the result as a new dashboard, and so on.
What these workflows all have in common, IMO, is that there are humans using the LLM as an aid to dive understanding, and then translating that understanding into more traditional, reliable tools. For this reason, I tend to think that the concept of autonomous "agents" is stupid, outside of a few very narrow contexts. That is to say, once you know what you want, you are generally better off with a reliable, predictable, LLM-free application, but LLMs are very useful in the prices of figuring out what you want. And MCPs are helpful there.
Da Vinci himself likely had dozens of nameless assistants laboring in his studio on new experiments with light and color, new chemistry, etc. Da Vinci was Da Vinci because of his vision and genius, not because of his dexterity with his hands.
Now there is a quote I should frame and hang in the company break room.
GoodRX and other discount providers generally work in one of two ways:
1) They have relationships with multiple PBMs, allowing you to choose the one who has negotiated the cheapest rate with the pharmacy for the drug in question. This is why it might be cheaper than your insurance: another PBM has negotiated a better deal.
2) The discounts come from patient assistance programs run by the manufacturers intended to reduce patient co-pays. Lately insurance companies have started to add clauses to their plans (called copay accumulators or copay maximizers) so that these discounts don't count as part of your copay or your deductible. So these types of discounts are going to be harder to get.
This all stems from a time when pharmacies were much less consolidated and vertically integrated than they are today.
One of the frustrations of the current system is that incentivizes sky-high drug prices. PBMs like high drug prices because they negotiate rebates (some of which they keep, but most they pay back to the insurer) and because the fees they charge to insurers are a percentage of the claims that go through. Pharmacies like high drug prices because they get more money paid them in reimbursements, and because the PBMs send them most of their customers. Manufacturers like high drug prices because they net more revenue, even if they later have to pay it back in the form of rebates, and in any case being on the formulary of major insurers is an existential issue for them. And insurers like high drug prices because they can max out patient co-pays, as the money returned to them in the form of rebates gets kicked into the general fund, thus allowing them to lower premiums, which is their primary axis of competition with other insurers.
The net effect is that you have sick people maxing out their deductibles in order to lower the premiums paid by healthy people--the exact opposite of how insurance is supposed to work. If I could wave a magic wand in Congress and make only a single surgical change to healthcare markets, the change I would make is banning rebates. They were anti-customer when John D Rockefeller used them to obtain a monopoly on oil, and they are anti-customer today.
A good place to read about these dynamics in American healthcare is drugchannels.net. The author is super well informed on how these plans are implemented.
Source: ran a startup targeting pharmacies (which failed) and currently work in a starup focused on discovering and developing new drugs.
The author rightly observes that despite the undeniable shift, women are still significantly underrepresented in STEM and therefore that cannot explain the lack of replication. There are still many other reasons besides innate differences that might explain it.
However, from the shareholder perspective, buybacks are more tax-efficient since those who sell their shares pay taxes on the gains, and those who hold end up with a larger share of the company and no tax.
This is all according to theory. In practice, there does also seem to be a sort of received wisdom among managers that buybacks are better for companies because they can cut them in times of trouble, whereas investors will perceive cuts in dividend as somehow foretelling bankruptcy.
I know you can accomplish a lot of this kind of direct content filtering with uBlock Origin, but as far as I'm aware you have to configure e.g. CSS selectors per site.