Drug design is hard because it’s a system problem not a molecule-protein interaction problem.
At one point the gold standard for generating leads was using combinatorial chemistry to experimentally screen through millions of compounds.
I’m not sure how wide-spread that is any more; it used a lot of reagents. And it ran into the real bottleneck: lead generation tells you little to nothing about the efficacy, toxicity, and side effects of a drug or biologic in actual humans.
In silicon will never do a better job at lead generation than combinatorial chemistry. It could help with the real bottleneck. But… there just isn’t a lot of high quality training data on this.
They’ve been pushing ML approaches to drug design since at least the early 2010s. The approach has evolved with the trends in ML. Overall, I think the results have been very mixed; much like the traditional pharma pipeline.
> At one point the gold standard for generating leads was using combinatorial chemistry to experimentally screen through millions of compounds.
I don't think this was ever the gold standard. I mean, some companies did employ these approaches, but they were generally viewed as low-quality, brute-force approaches (at least at the handful of companies I was associated with).
If you invent the ability to tragedise a particular commons, and you also do almost all of the work of turning the crank to make the commons tragic, and then you sit there benefiting from the tragedy of the commons while saying “we may be profiting from it, but the tragedy of the commons is not our problem”, eventually we will figure out what you’re up to.
Nothing these LLM companies are doing is comparable to a flying car. Perhaps we could ask whether a company making flying monkeys should care about the economic fortunes of bipedal monkeys.
Of course it is. It's a brand new technology with a great deal of potential and some caveats. Something like this has not existed in the history of humanity to date. It's truly mind boggling that a large chunk of humanity's corpus can be compressed to just a trillion bytes, then queried and recombined at whim. Heck, a flying car is less revolutionary.
What exactly would you want to regulate? And keep in mind that regulations are very effective at protecting incumbents, which, as of right now, also includes openAI.
What exactly would you be advocating for? That LLMs can't be used to generate mathematical proofs? It's not even a case of intellectual theft, academic papers are in the public domain or can be accessed through a subscription.
For my sanity I don't follow US politics too closely, if "super intelligence" is truly being pushed hard then it seems reasonable to assume the reason why is to assert control of truth. Who are you to disagree with Super Intelligence when it tells you that Haitians are eating dogs?
I’m happy he’s dumb. If he were smart, he’d be way better at corruption. His lowbrow corruption of scam meme coins and fleecing tech CEOs for a couple million is pretty benign versus what an actually smart person could do.
Yeah I wouldn’t be so happy about that. He’s that dumb because he was chosen for the dumbness and plausible deniability
There are much more incidious actors behind the scenes doing corruption at levels we can’t fathom whole Donald is the blunt force tool absorbing all the attention and negative press with his goofy corruption
No. His real innovation was shamelessness. A lot of high society functions on these "gentlemen agreements" where you can theoretically get away with murder but it'd be rather uncouth of you to do so. He doesn't care and it's led him very very far in life.
I definitely can't use the word "benign" given the immense damage to our culture, economy, and national standing he's done, but directionally agreed. We're fortunate he's an actual stupid person.
> as his brain melts he's lost most of his touch, but Trump does have a natural bullies intuitive understanding of how to use this type of dynamic. the winning move is to tell him to fuck off but that's not a battle a lot of people will want to pick with the president.
To shrug it off is just another way to avoid the confrontation, to make it seem not worth one's time. If the people destroying our future are imbeciles, what does this say about us?
You're underselling decades of preparation from conservatives trying to lead to exactly this point. Pretending like it's just a Trump issue is incredibly myopic. Trump didn't write Project 2025 and he couldn't implement it alone. Furthermore, we have a significant portion of the voting public who fully support this. The fact that it's Trump's 3rd election and second win should tell you there's something deeply wrong with Republican voters. Trump is just the latest symptom of the sickness of this country.
How could you possibly ever draw a line? It's Brandolini's law. It's flooding the zone. It's loyalty tests. It happens so frequently that there is no energy to focus on any one specific crime. It's all diffused under an umbrella of criminal shit. And each additional one is just one more thing thrown onto the pile no one will ever look into or prosecute because there's simply too much to dig into.
Right. No one believes he's smart enough to have engineered any of this, he's just a corrupt buffoon who was in the right place at the right time with the right amount of reality TV fame to benefit from all you just said.
That's the only bright point about all this. Trump is kind of a 1000 year perfect storm of fame, corruption, and psychopathy; and he's fast approaching his expiration date.
I don't expect any of us will live long enough to see a similar skid-mark on the underwear of humanity.
> Furthermore, we have a significant portion of the voting public who fully support this.
Define "significant" - Trump has the worst approval rating of any president ever measured. The "conservative coalition" was always going to Balkanize, anyone with a modicum of sense could see that, and they were unpopular before but now they're EXTREMELY unpopular.
I mean, "a conservative win" in a country where 1/3 of people have no religious preference, 1/5 marriages is multi-racial and acceptance of legalized gay marriage is in the 70's. Winning a battle isn't winning the war, they were deluding themselves into thinking they were going to "reshape" anything.
Now they've outed themselves as nutcases, but crucially, nutcases who aren't delivering on the economic promises that were cover for their fascistic impulses, in a coalition that was obviously going to implode once they reached power.
Maybe looking back, this will be seen as the "dead cat bounce" of Conservatism than anything else.
The way I broadly think of it is that the left has won the culture and the right has won the economy. But not for the entire right, just the ultra rich.
Zooming out of the intradecade view, all the power they have won't ever put gay and trans people back in the closet any more than they can reverse interracial marriage. And all the social media outrage about billionaires will do nothing to reverse wealth inequality.
Social media outrage, no. But a state which doesn’t apply laws equally and consistently opens itself up to violence, a fraction of which will be directed at the billionaire class.
We only need to eat one, the rest will fall in line.
I don’t think the old “right is good for the economy” canard holds up today. It seems to be a holdover from disciplined fiscal conservatism in the 1980s; 40 trillion in debt, $6 gas, and a new and unpopular war are quickly eroding that talking point.
Just shows how effective conservative propaganda is. They are uniformly terrible for the economy and it is proven literally every time they have power. Even by the 80's conservatives were fucking over the economy. Nixon probably represents the most significant shift leading us to where we are today (See Nixonomics). That's late 60's and early 70's which "coincidentally" follows the realignment of parties after the Democratic party passed the Civil Rights Act.
You have to remember exactly how few people are represented by current US policies. They are willing to elect increasingly fringe candidates, "Just to feel something", in the hopes that something will change for them.
Yes and knowledge of how terrible Leopold was is not uncommon, yes less well known than "Hitler was bad" but that's just about as close to a universally accepted truth as we've managed as a species, it's not a useful benchmark.
Mythos is what they thought was too dangerous to release, fable was what they made after they worked on cybersecurity detection. As they say in the notes, this version of sonnet now has a similar screening process
It sounds like you're thinking of cryptography, not general security.
The world uses many security products that have false positives and false negatives (firewalls, intrusion detections, wafs, fraud detection, spam...). Those aren't generally considered security through obscurity.
It's paired with rate limits, monitoring, account control, multiple classifiers, a deliberate safety margin, model design, and other things too.
(I think this was also why they faced the controversy over not having zdr in fable: they wanted to use logs to detect repeated attempts, etc. Possibly a bigger change with Opus/Sonnet is that it's zdr with the classifiers?)
The reason you see "dumb" refusals that "should clearly be allowed" is that they're using more traditional deterministic methods to deny prompts rather than just relying on the random LLM which you could bypass by luck.
The US also doesn't have free trade following the same criteria, labeling requirements differ by state, licensing regimes are different, the same kind of challenges that exist in Canada
Normally people have feelings about things before they are able to put them into words, I would imagine if you were asked a question like "what city would you most like to visit" then unless you've already thought about it a lot, then you would have to do substantial non-verbal thinking before you can come up with an answer, and once you have the answer you may respond "my favorite city is X" and you decided what X would be before you started the sentence.
Here's the best way I can explain how I "think" because it is mostly passive, but sometimes, I am actively thinking.
If I'm asked a hard question, like what city I'd most like to visit, I kind of have to look up cities and cross reference them with ones I've noted that I want to visit, or have visited and liked enough to go back, then actually read the list, and figure out which one, but each time I read a city, I might have a, this would be great in summer or winter or they have a cool festival in June. Which leads to additional question I have to ask. Which I think about roughly the same speed I speak, maybe a word or two faster.
My wife is the complete opposite. She can put together a sentence in an instant in "parallel" so when she says it, she though it already, but not as a sentence.
Human brains all work so crazily weird, and we are only just now started to realize all the nuances.
I have a transcript on my blog post. Someone copied the transcript here too, search "spice‑harvester" on this page (I asked it to replace some of my personal project names with words from the Dune universe).
The author's blog has what follows (link also follows):
[Seriously y'all in what universe would some "caveat" or another NOT "be a real one" by whatever severity you'd want to measure that AND/OR need of saying so ... ]
Claude (Original)
Force pushed. 1234567...890abcd main -> main (forced update).
Verified
Local main and origin/main both at 890abcd, in sync.
Every commit reachable from origin/main: no old string found.
Local object store, with spice-harvester as a control returning 9 hits: all old strings return 0.
Working tree clean, tests pass.
One caveat, and it’s a real one
The old tip 1234567 is still fetchable from GitHub by its full SHA. I confirmed this directly:
GitHub keeps unreachable objects until its own garbage collection runs, and there is no schedule you control. Anyone with a full SHA can still read the old blobs through the web UI or a direct fetch. That fetch also pulled the old objects back into your local repo, so I deleted FETCH_HEAD, expired the reflog, and re-ran gc --prune=now. The counts above are from after that cleanup.
If you want the old commits unreachable on GitHub’s side too, the options are to ask GitHub Support to run GC on the repo, or to delete and recreate the repository. Deleting and recreating is the only one you can do yourself. Say the word if you want help with either.
Nothing to commit this turn. The rewrite was the work, and the tree is clean.
Claude × Vomit GPT-OSS-20B
I force‑pushed the branch, updating main to commit 890abcd. I verified that my local main and origin/main are both at 890abcd. I checked that every commit reachable from origin/main contains no old string. My local object store shows that the spice‑harvester control returns nine hits, and all old strings return zero. The working tree is clean and the tests pass.
However, the old tip 1234567 is still reachable on GitHub by its full SHA. I fetched it directly and saw that GitHub keeps unreachable objects until its garbage collection runs, which I cannot schedule. Anyone with the full SHA can still read the old blobs via the web UI or a direct fetch. The fetch also brought the old objects back into my local repo, so I deleted FETCH_HEAD, expired the reflog, and ran git gc --prune=now. The counts I reported are after that cleanup.
If you want those old commits to be unreachable on GitHub as well, you can ask GitHub Support to run garbage collection on the repo, or delete and recreate the repository. Deleting and recreating is the only option you can do yourself. Let me know if you need help with either.
There is nothing to commit this turn. The rewrite was the work, and the tree is clean.
reply