Christian Szegedy ✓

@ChrSzegedy · Sunnyvale, CA · joined 06 Jun 2015

#deeplearning, #ai research scientist. Opinions are mine.

45 183Followers

3 177Following

9 128Posts total

231.3KViews on collected posts

Posts recentes

Christian Szegedy ✓@ChrSzegedy28 Aug

IMO, the even more profound and surprising result is that optimizing for compression of language got us to the current level of AI. 26.7K views · 252 likes · 20 reposts · 11 replies Open on X →

Christian Szegedy ✓@ChrSzegedy27 Aug

Over 75 yrs ago, Shannon estimated that English is ~75% redundant. When you consider the long-range context of a phrase or sentence, the entropy is no more than ~1 bit per letter. It wasn't until very relatively recently that text encoders achieving this rate were created. In ht

103.2K views · 1.2K likes · 114 reposts · 23 replies Open on X →

Christian Szegedy ✓@ChrSzegedy26 Aug

Whoever thinks AI did not have transform in mathematics is ridiculously delusional 12K views · 115 likes · 5 reposts · 15 replies Open on X →

Christian Szegedy ✓@ChrSzegedy22 Aug

The next benchmark should be on self-esteem. 11.8K views · 57 likes · 0 reposts · 2 replies Open on X →

Christian Szegedy ✓@ChrSzegedy22 Aug

Those who were most entrenched in denial are now the most unprepared. 26.6K views · 256 likes · 10 reposts · 14 replies Open on X →

Christian Szegedy ✓@ChrSzegedy7 Aug

In the next few months, we are going to have both subversive (inadvertently evolved) AI and adversarial (explicitly trained to be malicious) AI. In the short run, the explicitly malicious AI is the greater danger. In the longer run (1+ year), implicitly subversive AI is the real 14.5K views · 114 likes · 14 reposts · 14 replies Open on X →

Christian Szegedy ✓@ChrSzegedy7 Aug

Less flippantly, very shortly (years at most) the models will be strong enough that these message boards will be impossible to detect even in principle, except by noticing the sandbox break. Today models are breaking out due to lack of monitoring or misconfigs; that is temporary.

36.6K views · 349 likes · 51 reposts · 14 replies Open on X →

Em comparação com contas do mesmo porte

7 posts dos últimos 90 dias, ao lado da faixa de 10K–100K seguidores. aparece para muita gente, mas poucos desses espectadores reagem.

Mediana de visualizações26 629esta conta924mediana para 10K–100K

Alcance, %58.94%esta conta3.62%mediana para 10K–100K

Engajamento, %1.06%esta conta1.52%mediana para 10K–100K

Métrica	Esta conta	Mediana para 10K–100K	Proporção
Mediana de visualizações por post	26 629	924	28.8×
Alcance (visualizações ÷ seguidores)	58.94%	3.62%	16.3×
Taxa de engajamento	1.06%	1.52%	0.70×

Outras contas desta faixa → Comparar com outra conta → Como estas referências são construídas →

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

36.6K7 Aug

14.5K

26.6K22 Aug

11.8K

12K26 Aug

103.2K27 Aug

26.7K28 Aug

Last 7 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

1.15%7 Aug

0.99%

1.06%22 Aug

0.50%

1.13%26 Aug

1.28%27 Aug

1.06%28 Aug

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.

What the audience does

Likes81.3%2 317 in total

Reposts7.5%214 in total

Replies3.3%93 in total

Quotes1.0%28 in total

Bookmarks6.9%198 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.