Edward Raff

Sr. Director @CrowdStrike. Chair @CamlisOrg. Author of #InsideDeepLearning @ManningBooks & of JSAT Machine Learning library. PhD from & Visiting Prof @UMBC

[Edward Raff @EdwardRaffML]

I'm now officially a published book author @ManningBooks! Inside Deep Learning https://t.co/9bk8hZtenF ! Filling the need for a combination of practical "get something running" and understanding why things work and how the math relates to the code. @KirkDBorne for the forward!

[Edward Raff @EdwardRaffML]

Eg, I’ll ask “make sure this document has a linear reading flow where each section’s content is uniform in subject, and each depends only on previously stated results” and it fails spectacularly once you read.

[Edward Raff @EdwardRaffML]

I don’t know how field dependent it is, but I find Claude’s writing to be horrendous. I would have expected my field to be “easier” for this given highly technical nature, but so far 0 writing ability even in explicit requests.

[Mushtaq Bilal, PhD @MushtaqBilalPhD]

It seems like writing will no longer be an important part of the research process. It can easily be outsourced to AI.

What's going to matter is the way researchers curate their materials.

Claude Code was able to do a good job partially because I already had a well-curated list of materials, and I knew the direction of the argument.

That said, researchers will still need to learn writing because it helps you think through problems.

[Edward Raff @EdwardRaffML]

@FrnkNlsn URL doesn’t work

[Kristopher Micinski -- REBORN @krismicinski]

My second PhD student, Yihao Sun, will be doing his dissertation defense tomorrow at noon Eastern time. Yihao's work has been a true sight to behold. He has papers at ASPLOS, AAAI, NeurIPS, VLDB, among others. This Fall, he starts as faculty at Utah State University! (Zoom link.)

[Edward Raff @EdwardRaffML]

@yoshitomo_cs All of them. Never had an appeal work. I’ve seen reviewers call a paper “retarded”, rejecting a paper with all accept votes, enter in a meta review for what was clearly a different paper. Reject a paper with a AC note explaining in detail why the paper is being accepted. 0%

[Edward Raff @EdwardRaffML]

@avt_im I’d love to know how prescriptive you were on instructions vs giving goals. My experience thus far is that none of the llms can do performance sensitive coding without being extremely prescriptive

[Edward Raff @EdwardRaffML]

@PINguAR Congrats!

[Pinar Yanardag @PINguAR]

Thrilled to receive an NSF CAREER Award! Excited to push the frontiers of video generation toward controllable, cinematic tools for creating dynamic worlds. Grateful to my students, mentors, collaborators, and NSF. https://t.co/P358n6io6q

[Edward Raff @EdwardRaffML]

""CLAUDE HAS BAD CODE DESIGN ISNTINCS AND MUST SEEK USER GUIDANCE BEFORE EDITS UNDER ALL CIRCUMSTANCES OUTSIDE OF A PRE-MADE PLAN" lets see if this can stop it from being so stupid so quickly. Its annoying how helpful it is to get stuff done but the amount of stupid code...

[Edward Raff @EdwardRaffML]

@leland_mcinnes You might be familiar with mine from previous conversations but happy to make them explicit.

[Overleaf @overleaf]

Collaborating with non-LaTeX users just got a whole lot easier! 🤝 You can now import .docx (Word) files directly into Overleaf (we'll automatically convert them to LaTeX for you!), and export your work back to Word when you're done.

[Edward Raff @EdwardRaffML]

@tqchenml @cjnolet I say this means I’m owed an ice cream! 🍨

[Edward Raff @EdwardRaffML]

@krismicinski 2handy4me

[Edward Raff @EdwardRaffML]

Does anyone know if I'm missing something that I can't use @OpenAI 5.5 Pro from Codex? I much prefer Codex for research over the chat interface so that I can organize/clean/test/etc.

[Edward Raff @EdwardRaffML]

Oooph, on a streak of high reviewed papers getting rejected 😓 4x in the past month is super disappointing

[Edward Raff @EdwardRaffML]

@andy_pavlo @pvldb I’ve emailed like, 4 or 5 different editors now. I’ll just keep trying I guess 🫠

[Edward Raff @EdwardRaffML]

@pvldb is there anyone manning the emails who can answer questions for a first-time submitter trying to get their revision/rebuttal sorted? Like, the website is missing many details 😅

[Edward Raff @EdwardRaffML]

Claude needs a "persiveration" detector where it gets interrupted so an adult can tell it to stop being dumb and take some simple problem-solving steps.

[Edward Raff @EdwardRaffML]

Yea, stuff is cool. Yea, we are using it - I use it a good amount. But people gotta take a chill pill. 🧊💊

[Edward Raff @EdwardRaffML]

Can’t read without a subscription; but I’ll poke at the premise being an annoying “fantasy” for something we’ve seen play out many times before. CAD, 3d graphics, robotic automation, translation, and tons more often fail to really displace the now “outdated” approach …

[jasmine sun @jasminewsun]

Most people I know in AI think the median person is screwed, and they have no idea what to do about it.

I spent the last 3 months talking to dozens of researchers, economists, and policy experts about AI's impact on work; including reps from every frontier lab and several Congressional offices. Unfortunately, I was not reassured.

The AI industry is raising the alarm, but can't change course. These companies' core business model relies on the disruption they are warning about: their faith in full automation only makes them go faster.

Policymakers are waking up, but still paralyzed by data and debates. Econ wonks disagree on plenty, but even the limited scenario looks like a "painful transition" that will disempower millions of workers.

But an "underclass" is not inevitable, but rather a societal choice — and one we can and should stop. Instead of waiting for impact, we should start planning now to support workers through AI disruption. Whether policymakers can assuage concerns about economic security may determine if we get to reap AI's gains at all.

New from me for @NYTOpinion. I put a ton into researching what I think may be the biggest topic of the year, so hope you read it (gift link here!) https://t.co/NiGJpjyjzH

[Edward Raff @EdwardRaffML]

It’s not even like AI is hyper-amazing from a business perspective. It is expensive, error prone, and creates a naturally strained supplier/consumer dynamic. It’s super useful, it’s not magic. And we didn’t even touch the legal uncertainties on IP!