Alex S

540 posts

Alex S banner
Alex S

Alex S

@StecOverflow

Biologist ⏩ iOS Dev ⏩ ML Engineer ⏩ R&D ML for analytical chemistry. Hobbies: (dead) languages, paleontology 🇺🇦

Finland Se unió Temmuz 2015
764 Siguiendo142 Seguidores
Alex S retuiteado
Sam Santala
Sam Santala@SamSantala·
So am I reading this, right? @Adobe @Photoshop I can't use Photoshop unless I'm okay with you having full access to anything I create with it, INCLUDING NDA work?
Sam Santala tweet mediaSam Santala tweet mediaSam Santala tweet media
English
1.4K
10.2K
71.1K
12.2M
Alex S
Alex S@StecOverflow·
@alexchaomander What I find fascinating about this is that you guys were able to build the knowledge graph based on an especially contradictory dataset where different documents would claim the opposite things to be true. Also, seems like Obsidian is dead now. Maybe a lot of other things too.
English
0
0
0
50
Alex Chao
Alex Chao@alexchaomander·
🧐📈 Are knowledge graphs the key for more reliable RAG? Presenting GraphRAG: LLM-Derived Knowledge Graphs for RAG, a project from Microsoft Research applying graph machine learning to the indexing, retrieval, and querying of large private datasets. youtube.com/watch?v=r09tJf…
YouTube video
YouTube
English
4
6
19
2.2K
Alex S retuiteado
Bakhti Nishanov
Bakhti Nishanov@b_nishanov·
The only reason the Russian empire didn’t have a colony in Africa was because they tried and failed (in Djibouti). To be clear, Russia did have colonies all across Eurasia and N. America. It was a colonial power that never admitted its past, let alone reconcile with it.
Bakhti Nishanov tweet media
English
620
1.7K
10.1K
0
Alex S retuiteado
Arvind Narayanan
Arvind Narayanan@random_walker·
People are excited about using ChatGPT for learning. It's often very good. But the danger is that you can't tell when it's wrong unless you already know the answer. I tried some basic information security questions. In most cases the answers sounded plausible but were in fact BS.
Arvind Narayanan tweet media
English
54
279
1.5K
0
Alex S retuiteado
François Chollet
François Chollet@fchollet·
You can replace search engines with a LLM, you'll get a much better experience! (As long as you don't mind that most information you get is made up, and as long as you don't need to track sources, and...)
English
1
148
1.5K
0
Bojan Tunguz
Bojan Tunguz@tunguz·
The end is near.
Bojan Tunguz tweet media
English
44
372
2.8K
0
Alex S
Alex S@StecOverflow·
What are some good Bible verses to use in a sermon on strongly typed programming languages? #repent and #typesafety by #ChatGPT
Alex S tweet media
English
1
0
3
0
Alex S
Alex S@StecOverflow·
@tunguz The funny thing is, all the Bible references it put there are absolutely relevant, I’ve just checked. They are not random, those are the places it used as an “inspiration”.
English
2
0
34
0
Alex S
Alex S@StecOverflow·
@maxjaderberg  Explain Bayes' theorem in Beowulf style in Old English.
Alex S tweet media
English
0
0
4
0
Max Jaderberg
Max Jaderberg@maxjaderberg·
Rhyming Shakespearean paper summaries 🤯
Max Jaderberg tweet media
English
3
8
87
0
Alex S
Alex S@StecOverflow·
@Gerashchenko_en just a bluff to force Western countries to refrain from providing additional air defense systems to Ukraine.
English
0
0
0
0
Anton Gerashchenko
Anton Gerashchenko@Gerashchenko_en·
Attention, Germany, Britain, France, Norway, Sweden, Denmark! Solovyev is suggesting "starting careful work" on these countries, meaning bombings.
English
913
1.2K
3.8K
0
Alex S
Alex S@StecOverflow·
@ducha_aiki шо за сигнали з поверхні? 🤨
Українська
0
0
1
0
Dmytro Mishkin 🇺🇦
Dmytro Mishkin 🇺🇦@ducha_aiki·
ChatGPT can speak Ukrainian. But the quality is lower compared to English
Dmytro Mishkin 🇺🇦 tweet mediaDmytro Mishkin 🇺🇦 tweet media
English
2
0
4
0
Alex S retuiteado
Michael Black
Michael Black@Michael_J_Black·
I repeat: Easily produced science text that's wrong does not advance science, improve science productivity, or make science more accessible. I like research on LLMs but the blind belief in their goodness does a disservice to them and science. Here is an example from #ChatGPT 1/5
Michael Black tweet media
English
7
26
280
0
Alex S
Alex S@StecOverflow·
#ChatGPT doesn’t yet know what happened this year. Don’t tell it, let it be happy.
Alex S tweet media
English
0
0
1
0
Dmytro Mishkin 🇺🇦
Dmytro Mishkin 🇺🇦@ducha_aiki·
Don't ask ChatGPT chess advices. They look close to truth on the surface (and in the beginning), but don't make sense. 1st attempt is 1st screenshot, then I restarted and contunied the dialog.
Dmytro Mishkin 🇺🇦 tweet mediaDmytro Mishkin 🇺🇦 tweet mediaDmytro Mishkin 🇺🇦 tweet mediaDmytro Mishkin 🇺🇦 tweet media
English
3
1
14
0
Alex S
Alex S@StecOverflow·
@ducha_aiki it also can write poetry about that.
English
0
0
1
0
Alex S
Alex S@StecOverflow·
@nearcyan a lot of “hallucinations” in ChatGPT answers. It just insists on being right even when it’s obviously wrong.
English
0
1
4
0