Andrea Mocci 🇪🇺@🇨🇭

2.5K posts

Andrea Mocci 🇪🇺@🇨🇭 banner
Andrea Mocci 🇪🇺@🇨🇭

Andrea Mocci 🇪🇺@🇨🇭

@andreamocci

Find me on @[email protected] Junior Group Leader @ https://t.co/rWqucZlWrE . Tweets about software engineering, politics & more | He/him

Lugano, CH Katılım Şubat 2010
684 Takip Edilen436 Takipçiler
Alberto Martín López
Alberto Martín López@a_martinlopez·
Things that happened yesterday: - I got a paper accepted at @ICSEconf. - I submitted a paper to @issta_conf. - I submitted a ~1M€ project proposal to @snsf_ch. - I celebrated Halloween dressed as a banana. And all that with a broken wrist! Now, that was spooky!
Alberto Martín López tweet media
English
2
0
16
699
Andrea Mocci 🇪🇺@🇨🇭 retweetledi
Federico Yankelevich
Federico Yankelevich@yanke46·
1st Program Committee meeting for @VoxxedTicino (online). Stay tuned for the first (small) round of selection. 🤩
Federico Yankelevich tweet media
English
0
2
11
647
Andrea Mocci 🇪🇺@🇨🇭 retweetledi
Gary Marcus
Gary Marcus@GaryMarcus·
👇Superb new article from @apple AI: “we found no evidence of formal reasoning in language models . Their behavior is better explained by sophisticated pattern matching—so fragile, in fact, that changing names can alter results by ~10%!” 𝗧𝗵𝗲𝗿𝗲 𝗶𝘀 𝗷𝘂𝘀𝘁 𝗻𝗼 𝘄𝗮𝘆 𝗰𝗮𝗻 𝘆𝗼𝘂 𝗯𝘂𝗶𝗹𝗱 𝗿𝗲𝗹𝗶𝗮𝗯𝗹𝗲 𝗮𝗴𝗲𝗻𝘁𝘀 𝗼𝗻 𝘁𝗵𝗶𝘀 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻, where changing a word or two in irrelevant ways can give you a different answer. Strongly encourage you to read the whole thread.
Mehrdad Farajtabar@MFarajtabar

1/ Can Large Language Models (LLMs) truly reason? Or are they just sophisticated pattern matchers? In our latest preprint, we explore this key question through a large-scale study of both open-source like Llama, Phi, Gemma, and Mistral and leading closed models, including the recent OpenAI GPT-4o and o1-series. arxiv.org/pdf/2410.05229 Work done with @i_mirzadeh, @KeivanAlizadeh2, Hooman Shahrokhi, Samy Bengio, @OncelTuzel. #LLM #Reasoning #Mathematics #AGI #Research #Apple

English
97
577
2.3K
494K
Andrea Mocci 🇪🇺@🇨🇭
The wonders of AI. A popular & expensive vision model is predicting with 1.0 confidence that the language of the 1st line in red here is corse (okay), while the 2nd one is again - 1.0 confidence, with no alternatives - "bm", i.e., the Bambara language (en.wikipedia.org/wiki/Bambara_l…)
Andrea Mocci 🇪🇺@🇨🇭 tweet media
English
0
0
2
106
Andrea Mocci 🇪🇺@🇨🇭 retweetledi
Eric Torreborre
Eric Torreborre@etorreborre·
A no-code platform founded by Adam Chlipala, built on formal methods, functional programming and an AI interface: nectry.com/about.html. That's something to follow.
English
0
7
32
2.3K
Andrea Mocci 🇪🇺@🇨🇭 retweetledi
Joe Biden
Joe Biden@JoeBiden·
Joe Biden tweet media
ZXX
79K
158.7K
919.5K
446.7M
Andrea Mocci 🇪🇺@🇨🇭 retweetledi
Software Institute
Software Institute@usisoftware·
🏝️ The best summer #SIESTA is back on the seaside! Discover more about our software engineering summer school and register now! siesta.si.usi.ch/2024/
Software Institute tweet media
English
1
13
15
1.1K
Andrea Mocci 🇪🇺@🇨🇭 retweetledi
Scalar Conference
Scalar Conference@scalarconf·
Watch how to Minimize your Python Discomfort: A tour of NLP with Spacy, ScalaPy, and Scala 3 by @andreamocci during the Scalar Conference 2024 Subscribe to our YouTube channel so you don’t miss the rest of #scalarconf videos. youtube.com/watch?v=Z1dKpX…
YouTube video
YouTube
English
0
3
5
720
Andrea Mocci 🇪🇺@🇨🇭
Andrea Mocci 🇪🇺@🇨🇭@andreamocci·
@baldram @riccardo_cardin Yeah I just wanted to point out that particular issue of a by-name, the syntax is actually the same modulo even less-braces. I honestly wasn't 100% sure anymore that you could use a by-name param in an extension method but... obviously Scala is cool enough to allow it.
English
1
0
1
31
Riccardo Cardin
Riccardo Cardin@riccardo_cardin·
Do you have an old-fashioned #Scala function that still throws exceptions? With Raise4s, You can migrate to raising logic-typed errors using the '$catch' function. Don't worry; in version 0.0.3, the function will be renamed to 'catching' 😜
Riccardo Cardin tweet media
English
3
0
12
4K
Andrea Mocci 🇪🇺@🇨🇭
Andrea Mocci 🇪🇺@🇨🇭@andreamocci·
@baldram @riccardo_cardin I like it! I would first though get rid of the `() =>` by using a by-name param, and yes, it works with extension methods. In this way, the argument is not evaluated before the call. The same could be done with `$catch`, and I'm attaching an alternative syntax.
Andrea Mocci 🇪🇺@🇨🇭 tweet mediaAndrea Mocci 🇪🇺@🇨🇭 tweet media
English
4
0
3
81
@baldram@functional.cafe 🔸️ Marcin Szałomski
How about implementing an extension method and doing it straightforward as in the attached picture? The syntax will not only be more fluent but also more intuitive and easier to use. Btw. to get your example working, you probably had to run this in proper block e.g. `either`, or add a line like `given raise: Raise[UserNotFound] = new DefaultRaise`? Would you consider adding a working example like this in tests to document the use-case and functionality?
@baldram@functional.cafe 🔸️ Marcin Szałomski tweet media
English
2
0
5
260
Andrea Mocci 🇪🇺@🇨🇭 retweetledi
Martin Fowler
Martin Fowler@martinfowler·
Much of the chatter about GenAI is writing new code, but that's a small part of our work. Adam Tornhill talks to me and my Thoughtworks colleagues about how AI can help with refactoring. thoughtworks.com/insights/podca…
English
5
59
282
44.7K
Mario Fusco
Mario Fusco@mariofusco·
... and then you can traverse them using the new switch syntax #DevoxxGR
Mario Fusco tweet media
English
2
0
23
2K
Mario Fusco
Mario Fusco@mariofusco·
The combination of records and sealed classes is the Java way to define algebraic data types #DevoxxGR
Mario Fusco tweet media
English
5
17
128
7.9K
davidcshepherd
davidcshepherd@davidcshepherd·
@andreamocci @ShriramKMurthi @samth @guidosalva @ICSE2024 I remember that... what an intense town hall we had at ICSE that year! I think it was actually a good idea; maybe some variant of it will re-surface that can work to limit spamming submissions without causing too much outrage.
English
1
0
3
153