Arvind Narayanan Profile picture
Apr 12, 2019 11 tweets 6 min read Read on X
I used to share the popular view that creating and maintaining research software is pointless — no one will use it, you won’t get credit for it, and it will die once the lead author completes their PhD.

Here’s the story of how I changed my mind and what I learnt from it.
Seven years ago my team started a project to uncover creepy online tracking. We quickly realized we’d stumbled into a vast cesspit, and that we were going to be working on it for a while. We figured we’d save time by automating much of our workflow.
In retrospect, that was a good reason to build research software — do it if you think it will be worth your time even if no one except you and your collaborators ever use it. If other people do find it useful, that’s a bonus.
My team members @s_englehardt and Gunes Acar were/are the main developers of the tool. They are far better software engineers than I am. It ended up being stable & usable enough that others started using it without needing our help. This amazed me.
@s_englehardt I learnt that research software is more likely to succeed if (unlike me!) you’ve built software outside the research context. That's one reason why spending a couple of years doing software engineering gives you a leg up compared starting a CS PhD straight out of undergrad.
@s_englehardt It helped that we publicly released our software right away without waiting to collect our academic brownie points (a publication). By the time we wrote up the accompanying paper, there were many studies that used the tool, so we were able to present it as a mature system.
@s_englehardt Happily, we did get academic credit for it. We presented both the design of the software and our main findings in a 2016 paper (…), which has been cited a bunch and was selected by @futureofprivacy as one of the 2016 Privacy Papers for Policy Makers.
@s_englehardt @futureofprivacy I’m especially happy about the fact that even after @s_englehardt completed his PhD, the software continues to thrive. His new employer, Mozilla, has found it useful and has taken over its development…
@s_englehardt @futureofprivacy Meanwhile, over 30 studies have used the software….
Its use has expanded beyond studying online privacy: we are wrapping up a year-long investigation into “dark patterns”, led by @aruneshmathur, and I’m excited to share the results soon.
@s_englehardt @futureofprivacy @aruneshmathur In summary, I think building reusable research software is a totally underrated. I don’t mean to downplay the sweat and toil, but the payoffs are worth it, and the positive externalities make you feel good!
@s_englehardt @futureofprivacy @aruneshmathur P.S. Our creepiest findings using the tool are in our "No boundaries" series of blog posts, in which we explored how third-party scripts on websites have been extracting personal information in increasingly intrusive ways:…

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Arvind Narayanan

Arvind Narayanan Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @random_walker

May 16
In the late 1960s top airplane speeds were increasing dramatically. People assumed the trend would continue. Pan Am was pre-booking flights to the moon. But it turned out the trend was about to fall off a cliff.

I think it's the same thing with AI scaling — it's going to run out; the question is when. I think more likely than not, it already has.The image is a line graph titled "Top Airplane Speeds and Their Dates of Record, from Wright to Now," produced by the Mercatus Center at George Mason University. The graph tracks the progression of top airplane speeds from 1903 to around 2013. Here's a detailed description:  Y-Axis (Vertical Axis): Labeled "miles per hour (mph)," it ranges from 0 to 2,500 mph. X-Axis (Horizontal Axis): Labeled with years from 1903 to 2013 in increments of 10 years. Notable Annotations: Speed of Sound: Represented as a horizontal dashed line across the graph at approximately 760 mph. Reco...
By 1971, about a hundred thousand people had signed up for flights to the moon…
You may have heard that every exponential is a sigmoid in disguise. I'd say every exponential is at best a sigmoid in disguise. In some cases tech progress suddenly flatlines. A famous example is CPU clock speeds. (Ofc clockspeed is mostly pointless but pick your metric.)
Note y-axis log…Image
Read 11 tweets
Apr 30
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 🧵 This image is a scatter plot titled "Our simple baselines beat current top agents on HumanEval." It charts the performance of various computational models based on their human evaluation accuracy and cost. The horizontal axis represents cost, while the vertical axis shows human evaluation accuracy ranging from 0.70 to 1.00. Different models, such as GPT-3.5, GPT-4, and those from the Reflexion series, are plotted as points. The Pareto frontier, depicted by a dashed line, shows the most efficient trade-offs between cost and accuracy. Points are colored differently to indicate the c...

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 ()…
Here are the five key takeaways.…
AI agent accuracy measurements that don’t control for cost aren’t useful.  Pareto curves can help visualize the accuracy-cost tradeoff.  Current state-of-the-art agent architectures are complex and costly but no more accurate than extremely simple baseline agents that cost 50x less in some cases.  Proxies for cost such as parameter count are misleading if the goal is to identify the best system for a given task. We should directly measure dollar costs instead.  Published agent evaluations are difficult to reproduce because of a lack of standardization and questionable, undocumented evaluati...
Read 12 tweets
Apr 12
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.…
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.
Read 7 tweets
Dec 29, 2023
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…
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.

Screenshot from lawsuit: output from GPT-4 identical to actual text from NYT
Read 13 tweets
Aug 18, 2023
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.🧵…
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.
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.”…
Read 30 tweets
Jul 19, 2023
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…
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. Among those skeptical of the intentional degradation claim, the favored hypothesis for people’s subjective experience of worsening performance is this: when people use ChatGPT more, they start to notice more of its limitations.  But there is another possibility.  The user impact of behavior change and capability degradation can be very similar. Users tend to have specific workflows and prompting strategies that work well for their use cases. Given the nondeterministic nature of LLMs, it takes a lot of work to discover these st
Read 9 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!


0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy


3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!