ohad rozen (@ohadrozen) profile picture

PhD candidate for NLP and AI @BIU | Entrepreneur

Israel
Joined June 2009
Photo by ohad rozen (@ohadrozen)Photo by ohad rozen (@ohadrozen)Photo by ohad rozen (@ohadrozen)Photo by ohad rozen (@ohadrozen)Photo by ohad rozen (@ohadrozen)
"Replacing determinism ... with probabilistic code is fraught with edge cases..." Agreed! Probabilistic code fails because AI only accesses documented knowledge. Real-world ops rely on tribal knowledge- thousands of undocumented edge cases. Mapping that is key to making AI work.
Not nearly enough people are talking about the implications of Klarna rolling back some of their AI bets. Not knowing any of the details, I can guess why: Replacing determinism or humans with probabilistic code is fraught with edge cases and require new ways of software development and process engineering that aren't well solved yet. The implications to an entire generation of AI "apps" will be severe as more companies come to terms with the difficulty in getting products to work reliably in production with AI in the loop. Customer Service may be the first funeral signpost. The result will be that many startups will need to pivot to simply use AI for narrow use cases and otherwise act deterministically. So what have people funded then? An AI company? Not really. Just a really overpriced SaaS company at AI valuations.
51
โ€ข ื™ื•ื ื™ ื‘ืžื‘ื–ืง ื—ื™ืจื•ื! โ€ข
3
21
2
114
3,826
ื ืื•ื ืžื“ื”ื™ื ื‘ืื•ืจืš ื“ืงื” ืฉืžืกื‘ื™ืจ ื‘ืฆื•ืจื” ื”ืžื“ื•ื™ืงืช ื‘ื™ื•ืชืจ ืืช ื”ืกื˜ื ื“ืจื˜ ื”ื›ืคื•ืœ ื‘ื™ื—ืก ืœื™ืฉืจืืœ!
75
397
20
1,502
33,215
ื ืขืฉื” ืืช ื–ื” ืคืฉื•ื˜: 1. ืึธืกื•ึผืจ ืœื”ื›ื ื™ืก ืœืขื–ื” ืฉืื•ึผื ืกึดื™ึผื•ึผืขึท ื”ื•ึผืžึธื ึดื™ื˜ึธืจึดื™ 2. ืœึดืคึฐื ึตื™ ืฉื”ืฆืœื‘ ื”ืื“ื•ื ืžื‘ืงืจ ืืช ื›ึธึผืœ ื”ึทื—ึฒื˜ื•ึผืคึดื™ื 3. ืžืขื‘ื™ืจ ืœื”ื ืชึฐึผืจื•ึผืคื•ึนืช ื•ืชืž"ืœ ืœืชื™ื ื•ืงื•ืช 4. ืžืชืขื“ ื•ืžืฆืœื ืื•ืชื 5. ืžืขื‘ื™ืจ ื•ืœื•ืงื— ืžื›ืชื‘ื™ื (ืž/ืœื™ื•ื“ืขื™ ืงืจื•ื ื•ื›ืชื•ื‘๐Ÿ˜ฅ๐Ÿ˜ฅ) ื”ึดืคึฐืงึทืจึฐื ื•ึผ ืื•ึนืชึธื ื›ึฐึผื‘ึธืจ ืคึทึผืขึทื ืึทื—ึทืช...!!!
38
125
5
737
13,762
ืชื›ืœืก
25
176
10
1,016
19,125
ohad rozen retweeted
ืขื ื›ืœ ื”ื”ืชืคืชื—ื•ืช ื”ื˜ื›ื ื•ืœื•ื’ื™ืช ื›ื™ืฆื“ ื–ื” ืฉื”ืืœื’ื•ืจื™ืชื ืฉืœ ื˜ื•ื•ื™ื˜ืจ ืœื ื™ื•ื“ืข ืขื“ื™ื™ืŸ ืœื”ืขืœื™ื ืื•ื˜ื•ืžื˜ื™ืช ืจืื™ื•ืช ืœืคื™ ื”ืฆื•ืจืš?
ืžืžืฉืœื” ืฉืชืœื•ื™ื” ื‘ืžืคืœื’ื•ืช ื”ืขืจื‘ื™ื•ืช ื”ื™ื ืกื›ื ื” ืงื™ื•ืžื™ืช ื•ืžื™ื™ื“ื™ืช ืœื‘ื™ื˜ื—ื•ืŸ ื™ืฉืจืืœ. ื™ืฉ ืขื•ื“ ื–ืžืŸ ืœืขืฆื•ืจ ืืช ื”ื˜ื™ืจื•ืฃ ื”ื–ื”
24
70
6
653
We uploaded a new paper! Please check it out: arxiv.org/abs/2006.01067 Aligning Faithful Interpretations with their Social Attribution with @yoavgo What does it mean for an interpretation to be "faithful", or for a faithful interpretation to "explain decisions" to people?
2
15
1
44
ืžืืžืจ ื—ืฉื•ื‘ (ื•ืืจื•ืš) ืฉืžืกื‘ื™ืจ ืื™ืš ืืคืฉืจ ืœื ืฆื— ืืช ื”ืงื•ืจื•ื ื” ื•ืื™ืš ืคืขืœื• ืžื“ื™ื ื•ืช ืฉื•ื ื•ืช ื‘ืขื•ืœื. medium.com/@tomaspueyo/coronโ€ฆ ืกื™ื›ืžืชื™ ืื•ืชื• ื‘ืขื‘ืจื™ืช ื›ื“ื™ ืœื”ืงืœ: m.facebook.com/story.php?stoโ€ฆ
1
3
Overfitting in humans
1
2
12
Anyone currently writing a research paper who would be down to participate in a fun experiment? โ€“ at Brooklyn, NY
12
3
18
DandyLoop: A network of online shops who share visitors dandyloop.com via @maximesalomon on @producthunt producthunt.co/posts/2514
Thanks Peter. Great to hear your feedback! Would you like to have a quick call to see how we can increase your conversion?