Sander Dieleman ✓
@sedielem · London, England · joined 02 Dec 2014
Research Scientist at Google DeepMind (WaveNet, Nano Banana, Gemini Omni). I tweet about ML, music, generative models (personal account).
70 599Followers
2 045Following
2 814Posts total
271.4KViews on collected posts
Latest posts
Looking for new posts…
@sedielem @doomie I think the position is not posted for Zürich on the website
2.8K views · 5 likes
· 0 reposts · 2 replies
Open on X →
@sedielem That looks exciting, thanks for posting. Someone already asked about a referral, but I guess it can't hurt to have one?
1.2K views · 0 likes
· 0 reposts · 1 replies
Open on X →
We are hiring research scientists and engineers🧑🔬 on the Gemini Omni team! Both in the US (Mountain View / SF) and Europe (London / Zürich).
Apply here: https://t.co/8bNvJWQAAL https://t.co/sqHuJRoGbF
72.1K views · 587 likes
· 54 reposts · 9 replies
Open on X →
10 years today since we unveiled WaveNet!
Autoregression with long context before it was cool😅
Maybe it looks a bit silly now that we didn't use a Transformer, but we had a good reason: it would take another year for that to be invented🙃 Thanks @heiga_zen for the reminder! http
40.7K views · 501 likes
· 32 reposts · 22 replies
Open on X →
GIF
Recently, I had a great chat with my friend and former colleague @GarneloMarta, CSO at Fundamental (@LTMpredict).
We talked about diffusion models (duh 🤭), music generation, typicality, blogging, @kaggle and more (also restricted Boltzmann machines 🫣)
https://t.co/WFRrWI0Yhw ht
14.1K views · 191 likes
· 18 reposts · 6 replies
Open on X →
47:53
Cool to see people still remember this!
Nice reminder of a fun time at the Reservoir Lab back in 2015 with @317070, @avdnoord, @lpigou, right before a bunch of us graduated and joined @GoogleDeepMind. Perhaps unsurprising that a fair few of us work on generative media today😁
14K views · 87 likes
· 4 reposts · 2 replies
Open on X →
For the curious what I'm talking about, inceptionism was this thing by @zzznah @ch402 etal: https://t.co/npndN30GoE
Take an ImageNet model, fix output class and backprop to input, optimize input to maximize the output class and some smoothness auxloss. See img1.
Then back at htt
34.2K views · 69 likes
· 5 reposts · 3 replies
Open on X →
@sedielem i really just want a canonical version of this to be settled on and just work out of the box, it is so obviously the right idea for a lot of important use cases.
539 views · 3 likes
· 0 reposts · 1 replies
Open on X →
@sedielem @sedielem interesting to see it making a comeback. feels like we’re on the verge of unraveling some big potentials in language AI again.
307 views · 3 likes
· 0 reposts · 0 replies
Open on X →
@sedielem I think you touched on this topic in the conversations we had at m2l summer school last september, so I was really looking forward to reading more about it from an expert like you.
personally, I like that it turned into a historical account of language diffusion resear
688 views · 3 likes
· 0 reposts · 1 replies
Open on X →
New blog post: continuous diffusion for language is back!
This research direction receded into the background for a while, but as of this year, it is once again a hot topic. I wrote down a historical perspective and some thoughts on the recent revival.
https://t.co/gfDZgJblTu
89.4K views · 763 likes
· 135 reposts · 22 replies
Open on X →
Against accounts of the same size
12 posts from the last 90 days, next to the 10K–100K follower range. shown widely, but few of those viewers react.
Median views8 386this account993median for 10K–100K
Reach, %11.88%this account3.67%median for 10K–100K
Engagement, %0.70%this account1.61%median for 10K–100K
| Metric | This account | Median for 10K–100K | Ratio |
|---|---|---|---|
| Median views per post | 8 386 | 993 | 8.45× |
| Reach (views ÷ followers) | 11.88% | 3.67% | 3.24× |
| Engagement rate | 0.70% | 1.61% | 0.44× |
Others in this range → Compare with another account → How these benchmarks are built →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
89.4K24 Aug
688
307
53925 Aug
34.2K30 Aug
14K
14.1K7 Sep
40.7K8 Sep
72.1K14 Sep
1.4K
1.2K
2.8K
Last 12 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
1.03%24 Aug
0.58%
0.98%
0.74%25 Aug
0.23%30 Aug
0.67%
1.52%7 Sep
1.36%8 Sep
0.90%14 Sep
0.15%
0.09%
0.25%
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.61%.
What the audience does
Likes57.9%2 213 in total
Reposts6.5%248 in total
Replies1.8%70 in total
Bookmarks33.8%1 292 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Similar accounts
betway
@betway
141.2K followers
· 72.5K posts
FSSAI
@fssaiindia
141.2K followers
· 17.6K posts
Dimitri Revolt ✓
@reality_revolt_
141.2K followers
· 191 posts
(fan) Mario ✓
@Mario___RM
141.2K followers
· 57.7K posts
Roman Reigns SZN 💥 ✓
@reigns_era
141.2K followers
· 116.4K posts
Nathalie Iannetta
@nathiannetta
141.2K followers
· 22.8K posts
Conscious Lee ✓
@TheConsciousLee
141.2K followers
· 42.1K posts
とんかつ@インデックス投資 ✓
@tonkatsu_index
141.2K followers
· 30.1K posts