I think the point of Jev is to thread the needle of the gap between non-LLM classifiers and LLMs.
Classifiers like classical NNs require:
- annotated data, potentially a lot of it
- training
- inference
#2 and #3 aren’t a big deal if you have an ML engineer, but #1 will always be a potential headache no matter who you are. The tradeoff is that they could be quite fast, cheap, and you can get probabilities, not just classes.
With LLMs you get:
- zero shot classification (no dataset or training required)
- potentially can use third party model providers like OpenAI off the shelf. Don’t even need to host your own model.
The downside to LLMs is that they are comparatively slow and expensive to traditional classifiers. Historically they also were prone to hallucination or malformed responses, though not as much these days. You also can technically get log-probs back, but these aren’t equivalent to the classifier probabilities.
Jev gets you the zero-shot, zero-infra benefits of LLMs, while being closer to the speed and cost of traditional ML classifiers, as well as both classification and probability responses.
Early in my career I worked for a pharmaceutical giant as a statistician. One of the things we would work on was trying to determine how much the various advertising efforts the company did actually moved the needle. Every year was the same miserable experience. The company would spend millions and millions of dollars in various advertising channels. We would attempt to determine how effective it was, but the wayin which they advertised and the way we collected data made it very difficult to generate any kind of insight at all. We would make recommendations as to how to adjust execution moving forward to ensure we could reliably determine how much value we were getting out of these campaigns, all of which would be ignored. They would massage our findings to tell whatever story they needed and would do the same thing the next year.
The marketing division is its own organization, with their own incentives that don’t necessarily align with the broader company’s incentives. They also were heavily addicted to relying on external analytical consultants who could pump out all the slide decks and pie charts they wanted, which would inevitably show that the marketing was not only effective, but should be invested in more the following year.
The most interesting thing in the article to me is that there was a dual library/bookstore, and that it’s privately owned. I don’t think I’ve encountered a library that wasn’t publicly owned in the US.
They are common here. Membership is usually a dollar or two a month. Members store their books their on a kind of permanent loan so they don’t have to store so many books in their house. It’s unstaffed and just run by its members, and members replace any books they lose.
There is a private engineering library I visited in TN (anyone can drop in during business hours) that aims to preserve books that engineers used before everything became computer assisted.
None of the libraries in New York City are publicly owned. NYPL, BPL, and QPL are all private nonprofits that receive public funding. It’s an odd arrangement.
Claiming Takeo City Library is a private operation is factually incorrect because it is strictly "publicly owned" by the city government. In fact, the city owns the land and building, and pays the private company, Culture Convenience Club (CCC), a management fee from local tax revenues to run the public library. Conversely, CCC pays monthly rent to the city to operate the Starbucks and bookstore inside. Since the entire facility is publicly owned and funded by taxpayers, CCC has zero ownership and is merely a contracted operator.
It’s not shifting accountability. After WWII, France demanded that its colonial holdings, including French Indochina, be returned to it as a requirement for it joining NATO. The Vietnamese were not interested in returning to French colonial rule. This kicked off a decade-long war between the two that the French lost. The US had advisors to the southern Vietnamese during that time, and rather than fully pulling out, increased their presence as they tried to manage the void left by the French. This spiraled into the Vietnam War.
The US is on the hook for its decisions. But Europe doesn’t get off the hook for its own.
An equally cynical but opposite take is that data is what you turn to when you want to dispel with the mythology, hand-waving, and general bullshit that self-described “creatives” wrap themselves in in order to avoid admitting that they may not actually understand why something works or doesn’t work (and by extension, that their success may be dumb luck rather than talent).
I don’t think truth aligns with either extreme. Both are approaches that have their place. But if it was an “uncreative” with no ideas that simply took the time to look and discovered that the three point shot was absurdly more efficient than basically any other shot in basketball, what does that say about all the “creative” coaching geniuses who failed to discover it in half a century? And if basketball is now “less creative” because of it, that’s an indictment of the fact that the league refuses to address the underlying issue.
I think it’s telling that no one bemoans the fact that we design our medicine using data driven methods. I’m sure it would spice up our lives if the pharma companies dropped the boring data analysis and statistical methods and just slapped some compounds together through free-flowing creative passion!
Corruption does not mean bad policy. It’s government officials breaking the law to empower and enrich themselves. To the extent the items on your list even exist outside of the addled minds of conspiracy theorists, they aren’t corruption.
I could be misremembering but the Walking Dead season 1 and Game of Thrones season 1 came out around the same time, but didn’t really kill of any major characters until after the first season.
Both are based on pre-existing material but GoT source material predates the Walking Dead by 7 years.
reply