I get the impression that your takeaway is not to use LLMs, but I think your takeaway should be to give them more oversight. The way you describe this, it sounds like you went pure vibe coding.
> I asked the LLM to refactor for redundancy, quality, and size.
What did you expect when you asked it to make it smaller?
I’ve had good success with LLM refactoring. But the refactoring has never been “go make this better and smaller”. It’s “go do this specific thing to reduce redundancy” or maybe “give me a list of ways to improve the code redundancy” and then iterate with the LLM to improve the ideas and execute on the right ones.
I don’t think you can just vibe code your way out of a vibe coded mess.
No, my takeaway is the same as with working with any other developer.
If I take responsibility for the product, then I can't afford to accept sub-standard input; no matter the source. My main mistake was not immediately putting the kibosh on the LLM, as soon as it started thrashing. Instead, I played along, for waaaayyyy too long.
The request for refactoring was actually a fairly long and detailed prompt. I can't recall it exactly, but by that time, the context was well and truly poisoned. The LLM had decided that its way was best, and could not be persuaded, otherwise.
The same goes for humans, in a similar situation. I've encountered almost the exact same behavior, many times.
I am quite aware that my personal standards tend to be a lot tougher than what seems to be the norm, these days, so I am willing to admit that I may be demanding too much, but that's how I was trained.
But I have been making extensive use of LLMs, for the last few months, and they have been incredibly helpful. The Rubicon has been crossed. There's no going back, but I also need to make sure that I don't get drunk on Kool-Aid. LLMs are still very much a WiP.
The title is misleading. “Improvement in advanced Alzheimer’s disease following high-dose psilocybin” sounds like results from a broader study, not an anecdotal report.
There’s a reason the article includes “a case report” in the title.
It might be inappropriate for the target audience because they might mistake the treatment for the protocol, but it's not trying to be misleading I think. In any case the title got editorialised which I think settles this.
What’s “back then” to you? Flash grew up in the time of dial up when you could still get AOL install discs with 100 free hours in your typical grocery store PC magazine. I don’t recall people hating Flash a lot until later when it wasn’t a technical necessity anymore.
The first computer I remember using was a Compaq Portable with a green screen and DOS that my dad was allowed to bring home on weekends. I vividly remember going to Circuit City as a family to buy our first windows 3.1 machine.
Flash was very cool, at first, then it got used for WAY too much stuff that had no graceful degradation so you were stuck waiting a few minutes for an animation to load so you could see the content stuck behind flash.
Flash certainly became broadly hated. It had a pretty long stretch of being loved, and enabling content that was loved, though. Up until about 2005 or so, flash was critical tech for the young web. By 2010 it was clearly heading toward an end.
I remember strongly disliking Flash before 2005, and many tech-minded people I talked to agreed. It was an awful install, always needed an updated, was a memory hog, and was a pain in the ass to use without it crashing. Yes, it must have been great to create apps with (I never did), but it was not beloved. Flash games were beloved, not Flash.
I remember the time of browser plug-ins (not “extensions”). Everyone happily installed Flash, and the Crescendo midi plugin, and multiple other in-retrospect-ill-advised plugins to enable fun stuff to work in their browser.
The “everyone hates Flash” stuff came later. It served a purpose for quite a while and people loved it. Newgrounds was a place of magic.
I mean, flash was always a pain in the ass, even when you got it working. The animations and games were great and I'm a big fan of stuff that tries to make it easy to publish programs like that, but I was still a teenager when apple announced they weren't supporting it and I was genuinely happy because I was so annoyed using it even on a windows pc.
It was a fun and experimental time for sure. Way more stuff was weird in a good way. Standards hadn’t settled. All kinds of fun stuff was created in Flash that could not have been built with the standardized web tech of the day. I don’t really miss Flash but I do miss the early internet sometimes and Flash was part of that. (Remember when it was FutureSplash?)
I would be remiss if I didn’t post the most early-Internet-type thing I’ve encountered in a long time. Dungeon Soup.
It seems relevant here because the question was “How will this potentially help me if I get cancer?” and the answer is “Not at all unless you get a particular form of cancer that this applies to”.
> Bacteria are also all different, but still they are "one thing", and despite their diversity, antibiotics exist that can deal with many species of them at once.
Except people don’t ask “what if I get bacteria” the way they ask about cancer. If the story was about a new antibiotic that only affected 20% of common infectious bacteria strains and someone asked “in laypersons terms, how will this help me if I get a bacterial infection”, it would be appropriate to clarify that it only applies to some bacteria.
> Except people don’t ask “what if I get bacteria” the way they ask about cancer.
Yeah, but doctors also don't tell people "you have bacteria" or claim "we found a cure for bacteria". The lack of nuance on average is largely due to a lack of nuance from experts. The media treats cancer as one big thing and bacteria and viruses as separate things. Thus the average joe inherits 'treating cancer as one big thing' from the media.
I agree with you about the media. Cancer is often presented as a monolithic thing by the media. I don’t agree at all about experts. Doctors and scientists who research cancers do not lack nuance.
Oncologists are actually way more specific than even that. Because there are many forms of breast cancer and different treatments depending on the type.
But yeah, oncologists aren’t telling people “you have cancer” the way they might say “you have MRSA”.
Yeah, it's WAY more specific. We got a genetic breakdown, multiple pamphlets on the drugs being used, what they are targeting, and why they work (along with the risks).
Honestly, I think people probably get false impressions because cancer usually hits old people and old people are, frankly, often not reliable narrators.
> By the end of the project, we wrote nearly four times as many lines of test code as we wrote for the Swift interpreter itself.
This is the most interesting bit to me. Engineers consistently underestimate the amount of effort that testing demands for projects that need truly high quality, it’s nice to see this shared.
It’s real weird to see people argue that LLM output is no different than random gibberish and then handwave over the fact that it’s clearly not with terms like “training”, as if a steam of random garbage is trainable.
I quite literally created and productized predictive linguistics and behavioral vectors at Google.
If you had stopped to consider what I explained; you’d understand that it’s the process of turning random garbage into increasingly acceptable outputs.
Ie training the monkeys.
The insight you are missing is the rule of networked scale. It turns out that any reactive node scaled enough can form sophisticated predictive system given reward over a training topography, even if it starts out at garbage or is literally made of monkeys.
So it is garbage. And you can turn garbage into semi-intelligence.
A human child is born with no ability to speak intelligibly. All they can do is babble. Through years of training they gain the ability to speak intelligibly and communicate in advanced ways.
The act of successful training means it’s not garbage anymore.
> So it is garbage.
This statement is ultimately meaningless and I continue to find it weird that someone who works in this space would support this view. If you fundamentally change the nature of a thing, it’s no longer that original thing. Is tan HDD still random garbage after you fill it with family photos just because that’s how it starts?
If you start with a fire hose of literal sewage and install a series of filters culminating in a reverse osmosis step that pours clean drinking water out, the product is not shit even if the original input was.
I don’t believe that you can’t understand the distinction between “at one point this was garbage” and “at the present time this is still garbage”. You’re clearly smarter than that.
> I asked the LLM to refactor for redundancy, quality, and size.
What did you expect when you asked it to make it smaller?
I’ve had good success with LLM refactoring. But the refactoring has never been “go make this better and smaller”. It’s “go do this specific thing to reduce redundancy” or maybe “give me a list of ways to improve the code redundancy” and then iterate with the LLM to improve the ideas and execute on the right ones.
I don’t think you can just vibe code your way out of a vibe coded mess.
reply