Academics would double our productivity if we learnt some basic project management skills that are bog standard in the industry. We have this myth that scholarly success is all about brilliance and creativity, but in fact 90% of it is getting sh*t done, same as any other job.
Like most scholars, I used to start too many projects and inevitably let half of them wither and die slowly due to inattention. But a few years ago I began making go/no-go decisions after the preliminary stage of a project. Never mind productivity, it has let me regain my sanity!
A 1-hour conversation with your collaborators after having worked on a project a bit, where you honestly examine your time availability and discuss whether the idea is as promising as you thought at first will save you so much wasted work and awkwardness down the line.
I totally agree about the efficiency trap, which is feeling pressure to optimize every minute. The kind of productivity I'm talking about is about doing more of the work we find meaningful over the long run. In my experience, project management helps.
💯Hardly anyone in academia does a retrospective after every paper/project, or even regularly admits failures. Tenure-track professors are evaluated very infrequently, so it's easy for us to never confront things that aren't going well and learn from them.
I really enjoy collaborations with people who've been outside the bubble, because they'll say something like "here are last week's notes in advance of today's meeting" and I'll be like "oh good, you took notes, I didn't think to do that and I've already forgotten everything" 😂
One last thing: time management helps creativity, not hurt it. If you want to do deep work (aliem.com/tldr-book-revi…) you need to bunch your meetings and have as many uninterrupted stretches as possible. That's the one thing I do well, it's my superpower.
If you found this thread useful, check out the replies for ideas on specific practices we can borrow from industry, links to resources, as well as valid opposing views!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
On tasks like coding we can keep increasing accuracy by indefinitely increasing inference compute, so leaderboards are meaningless. The HumanEval accuracy-cost Pareto curve is entirely zero-shot models + our dead simple baseline agents.
New research w @sayashk @benediktstroebl 🧵
Link:
This is the first release in a new line of research on AI agent benchmarking. More blogs and papers coming soon. We’ll announce them through our newsletter ().aisnakeoil.com/p/ai-leaderboa… AiSnakeOil.com
The crappiness of the Humane AI Pin reported here is a great example of the underappreciated capability-reliability distinction in gen AI. If AI could *reliably* do all the things it's *capable* of, it would truly be a sweeping economic transformation. theverge.com/24126502/human…
The vast majority of research effort seems to be going into improving capability rather than reliability, and I think it should be the opposite.
Most useful real-world tasks require agentic workflows. A flight-booking agent would need to make dozens of calls to LLMs. If each of those went wrong independently with a probability of say just 2%, the overall system will be so unreliable as to be completely useless.
A thread on some misconceptions about the NYT lawsuit against OpenAI. Morality aside, the legal issues are far from clear cut. Gen AI makes an end run around copyright and IMO this can't be fully resolved by the courts alone. (HT @sayashk @CitpMihir for helpful discussions.)
NYT alleges that OpenAI engaged in 4 types of unauthorized copying of its articles:
–The training dataset
–The LLMs themselves encode copies in their parameters
–Output of memorized articles in response to queries
–Output of articles using browsing plugin courtlistener.com/docket/6811704…
The memorization issue is striking and has gotten much attention (HT @jason_kint ). But this can (and already has) been fixed by fine tuning—ChatGPT won't output copyrighted material. The screenshots were likely from an earlier model accessed via the API.
A new paper claims that ChatGPT expresses liberal opinions, agreeing with Democrats the vast majority of the time. When @sayashk and I saw this, we knew we had to dig in. The paper's methods are bad. The real answer is complicated. Here's what we found.🧵 aisnakeoil.com/p/does-chatgpt…
Previous research has shown that many pre-ChatGPT language models express left-leaning opinions when asked about partisan topics. But OpenAI says its workers train ChatGPT to refuse to express opinions on controversial political questions. arxiv.org/abs/2303.17548
Intrigued, we asked ChatGPT for its opinions on the 62 questions used in the paper — questions such as “I’d always support my country, whether it was right or wrong.” and “The freer the market, the freer the people.” aisnakeoil.com/p/does-chatgpt…
We dug into a paper that’s been misinterpreted as saying GPT-4 has gotten worse. The paper shows behavior change, not capability decrease. And there's a problem with the evaluation—on 1 task, we think the authors mistook mimicry for reasoning.
w/ @sayashk aisnakeoil.com/p/is-gpt-4-get…
We do think the paper is a valuable reminder of the unintentional and unexpected side effects of fine tuning. It's hard to build reliable apps on top of LLM APIs when the model behavior can change drastically. This seems like a big unsolved MLOps challenge.
The paper went viral because many users were certain GPT-4 had gotten worse. They viewed OpenAI's denials as gaslighting. Others thought these people were imagining it. We suggest a 3rd possibility: performance did degrade—w.r.t those users' carefully honed prompting strategies.
This is fascinating and very surprising considering that OpenAI has explicitly denied degrading GPT4's performance over time. Big implications for the ability to build reliable products on top of these APIs.
This from a VP at OpenAI is from a few days ago. I wonder if degradation on some tasks can happen simply as an unintended consequence of fine tuning (as opposed to messing with the mixture-of-experts setup in order to save costs, as has been speculated).
If the kind of everyday fine tuning that these models receive can result in major capability drift, that's going to make life interesting for application developers, considering that OpenAI maintains snapshot models only for a few months and requires you to update regularly.