/AI12h ago

Susan Zhang, who co-developed OPT-175B, warns ML engineers who only edit standard training scripts face automation within a year

She targeted workflows reliant on runs under 512 chips.

315911922157.1K
#47Shane Gu
Susan Zhang
Mahesh Sathiamoorthy
Original post
Susan Zhang@suchenzang#59inAI

if your bread-and-butter consists solely of:

- tuning hyperparams/config files

- fitting points on a log-log plot

- tweaking a few lines in http://model.py, http://transformer.py, http://optimizer.py, http://train.py

- waiting a week for <= 512 chips to free up and then another week for loss curves to converge

it is completely understandable to be stressed about becoming automated into irrelevance within the next year or so.

question is, do you wait for that to happen, or do you start doing something differently now?

6:00 AM · Jun 7, 2026 · 48.5K Views
Sentiment

Many users expressed distress at warnings that AI advances will automate routine ML engineering tasks, feeling personally attacked or facing professional decline while a few highlighted their broader skills.

Pos
10.0%
Neg
90.0%
11 comments with sentiment.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS7.9KBOOKMARKS8LIKES32REPLIES4
Mahesh Sathiamoorthy@madiator

We will be automating post training with post training.

Susan Zhang@suchenzang

if your bread-and-butter consists solely of:

- tuning hyperparams/config files

- fitting points on a log-log plot

- tweaking a few lines in http://model.py, http://transformer.py, http://optimizer.py, http://train.py

- waiting a week for <= 512 chips to free up and then another week for loss curves to converge

it is completely understandable to be stressed about becoming automated into irrelevance within the next year or so.

question is, do you wait for that to happen, or do you start doing something differently now?

10hViews 7.9KLikes 32Bookmarks 8
Elena Dyachkova@ElenaRunsNYC

@suchenzang This is most everyone in meta ads ranking ai 😭

11hViews 785Likes 7Bookmarks 1
Shane Gu@shaneguML

imo the metric isn't impact, it's impact / compute — however GPU-rich you are.

Susan Zhang@suchenzang

if your bread-and-butter consists solely of:

- tuning hyperparams/config files

- fitting points on a log-log plot

- tweaking a few lines in http://model.py, http://transformer.py, http://optimizer.py, http://train.py

- waiting a week for <= 512 chips to free up and then another week for loss curves to converge

it is completely understandable to be stressed about becoming automated into irrelevance within the next year or so.

question is, do you wait for that to happen, or do you start doing something differently now?

5hViews 610Likes 3Bookmarks 0
Aeron@aeronxbt

@madiator so the output is just. more post training input?

8hViews 147
Draven@notdrvx

@madiator double post training for that extra loop feel

is there a word for this or did we just make it up?

7hViews 121
Adam@aschiff19

@suchenzang When people let AI make their decisions for them, solutions continue to get built for yesterday’s problems…

here in the Triangle I find AI uptake to be at a slower clip, but companies here are focused on problem-founder fit and using AI as a tool, not racing against AI progress

9hViews 1.8K
Yash@yash1_

@suchenzang Just enjoying the " beginning of the end "

12hViews 794
RBA@benalt

@suchenzang Ouch.

9hViews 620
neon@neonnbits

@suchenzang I’m happy i do much more than this

11hViews 503
Patrick Baitman@taofanqq

@suchenzang Recursive self-improvement will fit those points better than u anon

7hViews 481
phil@big_algocracy

@suchenzang aggressively deepmind coded post (no shade)

7hViews 121Likes 1
Roman Vorushin@vorushin

@suchenzang No LLMs can drink Flat White as I can!

10hViews 249
gandamu@gandamu_ml

@suchenzang do it across several domains simultaneously so that more things magically make sense over time and it feels like progress. The delusion helps until I find out there's no edge because everyone else is doing the same thing.

4hViews 196
Shuying Luo@shuying_luo

@suchenzang When I see 512 chips, 👀that’s a luxury

7hViews 40
dani@absenteewarlord

@suchenzang the people who spend the whole next year pivoting will last maybe three months longer than the ones who coast. not worth it, just spend the severance time on vacation. who knows if you’ll even be alive in a year.

17mViews 18
Load more posts