Head of Engineering Branding, Wix, @WixEng | Author of Tiebreaker: The Power of Networking, Communities, and Personal Branding (Dec 2021)

Israel
Joined February 2008
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
For the first time, we're hosting a Claude Community meetup at @Wix Campus, TLV! Join the meetup on November 2 to examine the decisions behind building with AI: how to define a problem worth solving, judge whether your solution delivers real value, and turn successful experiments into reliable systems. Three talks with practical lessons: >> Amir Harel (Symphony by Wix) will show how to build the smallest useful version, evaluate it, and turn successful experiments into repeatable results. >> Guy Aga will walk you through building a real product and business, from market research and product development to launching an online store with Claude Code. >> Tali Despins will explore how to learn beyond your expertise, work through the discomfort of being a beginner, and combine skills across disciplines. ๐Ÿ“ Wix Campus, Tel Aviv ๐Ÿ“… November 2, 17:00 Limited spots available. Save your spot: luma.com/claude-9ja6
3
3
399
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
ืื™ืš ื—ื™ื–ื•ื™ ื”ืžื™ืœื” ื”ื‘ืื” ื”ืคืš ืœื‘ื™ื ื” ืžืœืื›ื•ืชื™ืช ืฉื™ื•ื“ืขืช ืœื›ืชื•ื‘ ืงื•ื“ ื•ืœื”ืคืขื™ืœ ืชื•ื›ื ื•ืช? ื’ืœืขื“ ื‘ืจืงืŸ ืจืืฉ ื’ื™ืœื“ืช ื”ื“ืื˜ื” ืกื™ื™ื ืก ื‘-Wix ื”ื ื’ื™ืฉ ืืช ื”ืžืืžืจ ื”ื—ื“ืฉ ืฉืœื• - From Autocomplete to Intelligence: How Language Became the Operating System of AI - ืœืžืกืžืš ืคืชื•ื— ื‘ืขื‘ืจื™ืช, ื‘ืฉื‘ื™ืœ ืฉื›ื•ืœื ื™ื•ื›ืœื• ืœืœืžื•ื“. ื”ืžืืžืจ ืฉืœ ื’ืœืขื“ ืžืกื‘ื™ืจ ืื™ืš ืžื•ื“ืœื™ ืฉืคื” ืœื•ืžื“ื™ื, ืœืžื” ื”ืฉืคื” ื ืžืฆืืช ื‘ืœื‘ ืžื”ืคื›ืช ื”-AI, ื•ืื™ืคื” ื”ื™ื›ื•ืœื•ืช ืฉืœื”ื ืขื“ื™ื™ืŸ ืžื•ื’ื‘ืœื•ืช. ืœื ืฆืจื™ืš ื™ื“ืข ืžื•ืงื“ื, ืจืง ืกืงืจื ื•ืช. ืžื•ื–ืžื ื™ื ืœืงืจื•ื, ืœืฉืื•ืœ ื•ืœื”ืฉืื™ืจ ื”ืขืจื•ืช ื‘ืžืกืžืš. docs.google.com/document/d/1โ€ฆ
AI can write code, use tools, and power autonomous agents. Yet at its core, it does one surprisingly simple thing: predict ๐Ÿ”ฎ the next word. So how did autocomplete become intelligence? In his new article, Gilad Barkan cuts through the weekly noise of new models and frameworks to explain why language became the operating system of AI. Learn how LLMs actually acquire intelligence, why Transformers and scale changed everything, and where todayโ€™s models still fall short. Along the way, he connects ideas from leading thinkers and AI pioneers, including @ylecun, @AndrewYNg, and @harari_yuval. No technical background required: wix.engineering/post/from-auโ€ฆ
7
8
2
125
27,345
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
Thanks @morad for making good things to happen @ @Wix ! I was happy to share my knowledge and experience from building AI tools (and Octocode), and I feel privileged to be in this position. For anyone who wants to learn and do the workshop, it's right here: github.com/bgauryy/octocode-โ€ฆ
Yesterday at @Wix Campus TLV, we got hands-on with Octocode ๐Ÿ™ in a practical workshop led by @guy_bary. We built an MCP server, created tools AI agents can use, and practiced researching real codebases. We also explored how Octocode connects local code, GitHub, and npm to help agents ground their answers in evidence, alongside best practices for developing reliable tools with MCP and Skills. This is an important โค๏ธ part of what we do at Wix Engineering: sharing practical knowledge with the developer community and the wider industry. Workshops like this give developers room to experiment, ask questions, and bring what they learn back to their teams - helping them build better and move faster. Thank you, Guy, and everyone who joined us to learn and build together! * Slides here >> github.com/bgauryy/octocode-โ€ฆ * Octocode here >> octocode.ai/
2
4
210
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
ื”-AI ืžืื™ืฅ ืืช ื›ืชื™ื‘ืช ื”ืงื•ื“. ืื‘ืœ ื›ืฉื”ืชื›ื ื•ืŸ, ื”ื‘ื“ื™ืงื•ืช ื•ื”-Review ืœื ืขื•ืžื“ื™ื ื‘ืงืฆื‘ - ืฆื•ื•ืืจ ื”ื‘ืงื‘ื•ืง ืคืฉื•ื˜ ืขื•ื‘ืจ ืœืžืงื•ื ืื—ืจ. ื‘ืžืกื’ืจืช ื”ืžื™ื˜ืืคื™ื ืฉืœื ื• ืœืงื”ื™ืœืช ื”ืžืคืชื—ื™ื, ืฉื™ืชืคื ื• ื‘ื›ืžื” ืชื•ื‘ื ื•ืช ื˜ื›ื ื™ื•ืช ืงืจื™ื˜ื™ื•ืช ืžื”ืขื‘ื•ื“ื” ืขื ื‘ื™ื ื” ืžืœืื›ื•ืชื™ืช. ืขื›ืฉื™ื• ืจื™ื›ื–ื ื• ืื•ืชืŸ ื‘ื›ืชื‘ื” ื‘ื’ื™ืงื˜ื™ื™ื - @geektimecoil. ๐Ÿงฐ ืื™ืš ื‘ื•ื ื™ื ืชืฉืชื™ืช ืงื•ื ื˜ืงืกื˜ ืฉืžื—ื‘ืจืช ืกื•ื›ื ื™ื ืœืงื•ื“, ืœืชื™ืขื•ื“ ื•ืœืžื™ื“ืข ืžืคืจื•ื“ืงืฉืŸ, ื•ื–ื” ืžื‘ืœื™ ืœื”ืขืžื™ืก ื”ื›ื•ืœ ืขืœ ื”ืคืจื•ืžืคื˜ ืื• ืขืœ AGENTS .md. โš™๏ธ ืื™ืš ืžืขืจื™ื›ื™ื ืกื•ื›ืŸ ืฉืœื (ื›ื•ืœืœ ื”ื›ืœื™ื ื•ืกื•ื›ื ื™ ื”ืžืฉื ื” ืฉืœื•), ื•ืžื–ื”ื™ื ืจื’ืจืกื™ื•ืช ื•ืฉื™ื ื•ื™ื™ื ื‘ืขืœื•ื™ื•ืช ื•ื‘ื–ืžื ื™ ื”ืจื™ืฆื”. ๐Ÿ› ๏ธ ืื™ืš ืžื—ื‘ืจื™ื ืกื•ื›ื ื™ื ืœืœื•ืœืืช ื”ืžืฉื•ื‘ ืฉืœ ื”-CI, ื›ื“ื™ ืฉื™ื•ื›ืœื• ืœื ืชื— ื‘ืขื™ื•ืช, ืœืชืงืŸ ื•ืœื”ืจื™ืฅ ื‘ื“ื™ืงื•ืช ืžื—ื“ืฉ. ๐Ÿงฐ ื•ืื™ืš ืžื•ื“ืœ ื”-xEngineer ืžืฉืœื‘ ืžื•ืžื—ื™ื•ืช ื•ื”ื‘ื ื” ืขืžื•ืงื” ืขื ืื—ืจื™ื•ืช ืžืงืฆื” ืœืงืฆื” ื’ื ืขืœ ืืจื›ื™ื˜ืงื˜ื•ืจื”, ื•ื’ื ืขืœ ืื™ื›ื•ืช ื”ืชื•ืฆืื” ื‘ืคืจื•ื“ืงืฉืŸ. ืชืฉืขื” ืžืคืชื—ื™ื ื•ืžืคืชื—ื•ืช, ืืจื‘ืข ื–ื•ื•ื™ื•ืช ื—ืฉื•ื‘ื•ืช ื•ืฉืืœื” ืื—ืช ืงืจื™ื˜ื™ืช: ืื™ืš ื”ื•ืคื›ื™ื ืงื•ื“ ืฉื ื›ืชื‘ ืžื”ืจ ื™ื•ืชืจ ืœืžื•ืฆืจื™ื ืฉืžื’ื™ืขื™ื ืœืคืจื•ื“ืงืฉืŸ ืžื”ืจ ื™ื•ืชืจ - ืžื‘ืœื™ ืœื”ืชืคืฉืจ ืขืœ ืื™ื›ื•ืช? ืœื›ืชื‘ื” ื”ืžืœืื”: geektime.co.il/how-ai-is-chaโ€ฆ ืขื: @aviranm, @Asafy3, @guy_bary, @snirshr
1
5
1
3
935
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
From building production AI agents to hiring the engineers who will build what comes next >>> we're proud of @AshkenaziOrtal and Guy Nahum, who are bringing their experience to @Reversim Summit 2026! ๐Ÿ‘ ๐Ÿ‘ ๐ŸŽค Ortal will explore what happens when one LLM skill becomes hundreds: how to define skills, catch common failures, and route requests to the right capability in a system handling 300K conversations a month. Her talk builds on research co-authored with Vitalii Kloz and Mykhailo Ulianchenko, published on @arXiv. ๐ŸŽค Guy will tackle "Hiring Post-Claude": what makes a great developer today, why junior engineers matter, and how @Wix is rethinking hiring and development programs to cultivate the #xEngineer. Heading to #Reversim? Don't miss their sessions!
2
6
522
ื—ื˜ืคื ื• ืžื›ื•ืช ื‘ื’ืœืœ ื”ื“ืขื” ืฉืœื ื•
158
454
45
2,667
51,595
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
This research by @AshkenaziOrtal, Vitalii Kloz, and Mykhailo Ulianchenko, published on @arxiv - will be featured at the upcoming @reversim 2026 conference! ๐Ÿ‘ Don't miss Ortal's session, "From a Single Skill to a Sea of Skills: Architecting LLM Agents in Production". summit2026.reversim.com/sessโ€ฆ
Really impressive work from @Wix data science team, who published research based on a real challenge they saw and solved in our own customer care agent. They researched how to help AI agents distinguish between skills that are relevant to a userโ€™s question and those that can actually be executed for that specific user. This helps users get more relevant and useful support from the start. The team was able to find a more precise approach that reduced the information the model needed to process by 90.5% and saved 228.8 million tokens. A huge achievement. Congratulations to @AshkenaziOrtal, Vitalii Kloz, Mykhailo Ulianchenko, and the entire team.
3
3
8
554
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
1,000+ MySQL servers. 160+ clusters. Multiple regions. Zero downtime. Join us on October 28 as Shmuel Mekonen and @Danaavital58 break down how they pulled it off - the architecture, automation, tradeoffs, failures, and key decisions: wix.engineering/migrations-aโ€ฆ
1
1
1
336
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
ืœืกื•ื›ื ื™ AI ื›ื‘ืจ ื™ืฉ ื’ื™ืฉื” ืœืžื™ื™ืœื™ื, ืœืงื•ื“, ืœื›ืกืฃ ื•ืœืžืขืจื›ื•ืช ืืžื™ืชื™ื•ืช. ืื‘ืœ ืื– ืื—ื“ ืžื”ื ื—ื•ืฆื” ืืช ื”ื’ื‘ื•ืœ, ื˜ื•ืขื” ืื• ืžื‘ืฆืข ืคืขื•ืœื” ืฉืœื ื‘ื™ืงืฉื ื• - ื•ืžืฉืื™ืจ ืื—ืจื™ื• ื ื–ืง. ืžื™ ืื—ืจืื™ ืขืœ ื–ื”? ื‘ื“ื™ื•ืง ื‘ืฉืืœื” ื”ื–ื• ื”-FTC, ืจืฉื•ืช ื”ืกื—ืจ ื”ืคื“ืจืœื™ืช ืฉืœ ืืจื”ืดื‘, ืžืชื—ื™ืœื” ืขื›ืฉื™ื• ืœื”ืชืขืจื‘. ื”ืจืฉื•ืช ืคืชื—ื” ื‘ื—ืงื™ืจื” ื ื’ื“ OpenAI, ืื ืชืจื•ืคื™ืง ื•ื—ื‘ืจื•ืช AI ื ื•ืกืคื•ืช ืกื‘ื™ื‘ ื”ืกื™ื›ื•ื ื™ื ืฉืœ ืกื•ื›ื ื™ื ืื•ื˜ื•ื ื•ืžื™ื™ื. ืœืคื™ ืจื•ื™ื˜ืจืก, ื–ื• ืคืขื•ืœืช ื”ืื›ื™ืคื” ื”ืคื“ืจืœื™ืช ื”ืจืืฉื•ื ื” ื‘ืืจื”ืดื‘ ืฉืžืชืžืงื“ืช ื™ืฉื™ืจื•ืช ื‘-ืดrogue AI agentsืด - ืกื•ื›ื ื™ื ืฉื—ื•ืจื’ื™ื ืžื”ื’ื‘ื•ืœื•ืช ืฉื”ื•ื’ื“ืจื• ืœื”ื. ื”ืžืฉืžืขื•ืช ืคื” ื”ื™ื ื’ื“ื•ืœื” ืžืื“. ืื ืขื“ ืขื›ืฉื™ื• ื”ื‘ื˜ื™ื—ื•ืช ืฉืœ ืกื•ื›ื ื™ AI ื”ื™ืชื” ื‘ืขื™ืงืจ ื‘ืขื™ื” ื”ื ื“ืกื™ืช. ืขื›ืฉื™ื• ื”ื™ื ืžืชื—ื™ืœื” ืœื”ืคื•ืš ื’ื ืœืฉืืœื” ืฉืœ ืื—ืจื™ื•ืช ืžืฉืคื˜ื™ืช ื•ืื›ื™ืคื”. ื›ื›ืœ ืฉืื ื—ื ื• ื ื•ืชื ื™ื ืœืกื•ื›ื ื™ AI ื™ื•ืชืจ ื›ื•ื— ืœืคืขื•ืœ ืœื‘ื“, ืœื’ืœื•ืฉ, ืœื”ืจื™ืฅ ืงื•ื“ ื•ืœื”ืคืขื™ืœ ืžืขืจื›ื•ืช, ื”ืฉืืœื” ื›ื‘ืจ ืœื ื ืฉืืจืช ืจืง ืดื›ืžื” ื”ื ื—ื›ืžื™ืืด, ืืœื ืžื™ ืื—ืจืื™ ื›ืฉื”ื ื˜ื•ืขื™ื.
4
4
2
19
1,533
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
Bae Code can now work on your own enterprise GitHub repo in the cloud, just like you would work on Base44. Try it out.
Today weโ€™re introducing Base Code. Itโ€™s an early preview of how we think software will be built in the future. Itโ€™s a product that encapsulates everything weโ€™ve learnt from: - How our users are building software. - How weโ€™re building internally, scaling to hundreds of millions of dollars in revenue while keeping our engineering team very small and focused. Here are some of the principles behind it, which align with how we at @Base44 think about the future of software engineering: Cloud: - Software will move past local desktop environments (where current tools are widely used) and move to the cloud. - Moving everything to the cloud is not easy. Setting up dev environments with databases, infra components, services, mock data, etc. is easier said than done. - Base Code first scans your code repo and sets up everything - every infra component (databases, Redis, etc.) - in the cloud. It runs nonstop until your preview environment is ready. - From an enterprise standpoint, itโ€™s also the logical thing to do: a centralized, governed dev environment instead of handing out keys and secrets to all team members. - Once Base Code does that, EVERY team member can work in this environment. No more syncing local environments. Internally, this (cloud) is one of the main things that enabled us to move so fast. Collaboration: - Once everything is in the cloud, everybody can write software from anywhere (any browser, your phone, WhatsApp / iMessage coming soon). - You can easily see what everyone is working on, where theyโ€™re at, and how theyโ€™re prompting - and jump to their environment in one click. - Weโ€™ve built many great collaboration features from the ground up. Loops, automations, software factories: - A full, working cloud environment allows for many advanced capabilities we will unveil soon-think agents running in the browser and testing on every device and in any browser. - It also allows the use of automations to build software factories-e.g., agents reading support tickets, identifying bugs to fix or features to develop, implementing the changes, verifying them in the cloud, and potentially pushing them. And lastly, thereโ€™s a real advantage in being model agnostic. -------------- As always, weโ€™re releasing it very early. Itโ€™s far from perfect-but weโ€™re looking for early feedback so we can build it together with our great community and make this vision a reality. *Weโ€™re giving it away for free for 30 days*. Give Base Code a try. Connect any GitHub repo, get your live preview, and share it with your team. Iโ€™d love to hear what works and what doesnโ€™t: app.base44.com/base-code
1
3
7
675
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
ืกืคื™ื™ืกืืงืก ๐Ÿš€ ืขืฉืชื” ื”ื™ื•ื ืฆืขื“ ื—ืฉื•ื‘ ืงื“ื™ืžื”: ืžืฉื—ืจืจืช ืœื•ื•ื™ื™ื ื™ Starlink V3 ืžื”ื“ื•ืจ ื”ื—ื“ืฉ ื‘ืขื–ืจืช ื”ืกื˜ืืจืฉื™ืค. ืœืžื” ื–ื” ื—ืฉื•ื‘? ื›ื™ ืœืคื™ SpaceX, ืฉื™ื’ื•ืจ ืื—ื“ ืฉืœ ืกื˜ืืจืฉื™ืค ื™ื›ื•ืœ ืœืงื—ืช ืื™ืชื• ื›-60 ืœื•ื•ื™ื™ื ื™ V3. ื›ืœื•ืžืจ ืฉื™ื’ื•ืจ ืื—ื“ ืฉืœ Starship ื™ื•ื›ืœ ืœื”ื•ืกื™ืฃ ืœืจืฉืช ืงื™ื‘ื•ืœืช ืฉืฉื•ื•ื” ื‘ืขืจืš ืœื›-20 ืฉื™ื’ื•ืจื™ Falcon 9 ืขื ืœื•ื•ื™ื™ื ื™ V2. ืฉื™ืคื•ืจ ืื“ื™ืจ ื‘ื™ืขื™ืœื•ืช ื”ืคืจื™ืกื” ืฉืœ Starlink.
3
1
1
49
4,053
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
ืžื˜ื ื—ื•ืฉืคืช ืžืฉืงืคื™ VR ื—ื“ืฉื™ื ื‘ืžืฉืงืœ 100 ื’ืจื ื‘ืœื‘ื“ - ื•ื™ืฉ ืขื•ื“ (ืคืœื•ืก ืขื“ื›ื•ืŸ ื—ืฉื•ื‘ ๐Ÿ‡ฎ๐Ÿ‡ฑ ืขืœ ื™ืฉืจืืœ!) ื‘ื›ื ืก Meta Connect 2026 ืžื˜ื ื”ืฆื™ื’ื” ืืช Meta VR Glasses, ืžืฉืงืคื™ VR ืงืœื™ื ืžืื“, ื›ืฉื”ืžืขื‘ื“ ื•ื”ืกื•ืœืœื” ื ืžืฆืื™ื ื‘ื™ื—ื™ื“ืช ืžื—ืฉื•ื‘ ื ืคืจื“ืช ื‘ื›ื™ืก.ย  ื”ื ื›ื•ืœืœื™ื ืžืกื›ื™ micro-OLED ื‘ืจื–ื•ืœื•ืฆื™ื™ืช 5K, ืฉืœื™ื˜ื” ื‘ืืžืฆืขื•ืช ื”ืขื™ื ื™ื™ื ื•ื”ื™ื“ื™ื™ื. ื”ืžื—ื™ืจ: 1,299 ื“ื•ืœืจ. * ื”ืขืจื”: ืœื ื™ื•ื“ืข ืื ื–ื• ืชื”ื™ื” ื”ืฆืœื—ื”, ืื‘ืœ ื–ื” ื‘ื“ื™ื•ืง ืžื” ืฉืืคืœ ื”ื™ื• ืฆืจื™ื›ื™ื ืœื”ืฆื™ื’. ืคืกืคื•ืก ืฉืœ ืืคืœ. ื‘ืžืงื‘ื™ืœ ื”ื ื”ื›ืจื™ื–ื• ืขืœย Ray-Ban Meta Audioย ืžืฉืงืคื™ื™ื ืœืœื ืžืฆืœืžื” (*ืฉื•ื‘ ื”ืขืจื”: ืื•ืœื™ ื›ื”ืžืฉืš ื™ืฉื™ืจ ืœื‘ื™ืงื•ืจืช ืขืœ ืคื’ื™ืขื” ืงืฉื” ื‘ืคืจื˜ื™ื•ืช), ื•ื“ื•ืจ ื—ื“ืฉ ืฉืœ Ray-Ban Meta ืขื ืฆื™ืœื•ื 3K ื•ืกื•ืœืœื” ืžืฉื•ืคืจืช. ืกื•ื›ืŸ ื”-AI ืฉืœ ืžื˜ื Muse ืžืงื‘ืœ ืฉื™ื—ื•ืช ืงื•ืœื™ื•ืช, ื‘ื™ืฆื•ืข ืžืฉื™ืžื•ืช ื‘ืจืงืข, ืื™ื ื˜ื’ืจืฆื™ื•ืช ืขื ืฉื™ืจื•ืชื™ ืฆื“ ื’ืณ - ื•ื’ื ื™ื›ื•ืœืช ืœืขื‘ื•ื“ ื“ืจืš ื”ืžืฉืงืคื™ื™ื. ืžื˜ื ื’ื ื”ืฆื™ื’ื” ืืช Muse Charmย - ืžื›ืฉื™ืจ AI ืงื˜ืŸ ืœื›ื™ืก ืฉืžืืคืฉืจ ืœื“ื‘ืจ ื™ืฉื™ืจื•ืช ืขื ื”ืกื•ื›ืŸ ื‘ืœื™ ืœื”ื•ืฆื™ื ืืช ื”ื˜ืœืคื•ืŸ. (* ืฉื•ื‘ ื”ืขืจื”: ื–ื” ืžื” ืฉ OpenAI ื™ืฆื™ื’ื• ื‘ืงืจื•ื‘. ืžื˜ื ื”ื–ื“ืจื–ื” ื•ืขืงืคื” ืื•ืชื) ื•ื™ืฉ ื’ื ื”ืงืฉืจ ื—ืฉื•ื‘ ๐Ÿ‡ฎ๐Ÿ‡ฑ ืœื™ืฉืจืืœ: ืžืฉืงืคื™ Ray-Ban Meta ื•-Oakley Meta ื™ื’ื™ืขื• ืจืฉืžื™ืช ืœื™ืฉืจืืœ - ืขื•ื“ ื‘ 2026. ื–ื” ืขื•ื“ ืจื’ืข. ืœืื•ืจ ืขื ื™ื™ืŸ ื”ืคืจื˜ื™ื•ืช - ื”ืจืฉื•ื™ื•ืช ื—ื™ื™ื‘ื•ืช ืœื”ื›ื™ืจ ืืช ื–ื”.
5
2
1
25
26,028
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
ืดื™ืฉ ืœื™ ืขื•ื“ ื“ื‘ืจ ืื—ื“ ืœื”ืฆื™ื’ ืœื›ืืด (ืขืœ ืžืฉืงืœ Oh, one more thing ืฉืœ ื’ืณื•ื‘ืก) ืฆื•ืงืจื‘ืจื’ ื’ื ื”ืฆื™ื’ ืืช Muse Charm: ืžื›ืฉื™ืจ AI ื–ืขื™ืจ, ืžืขื™ืŸ ื˜ืžืื’ื•ืฆื™ ืฉืืžื•ืจ ืœืชืช ื’ื™ืฉื” ืžื”ื™ืจื” ืœืกื•ื›ืŸ ื”ืื™ืฉื™ Muse ืžื‘ืœื™ ืœื’ืฉืช ืœื˜ืœืคื•ืŸ. ื”ืžื›ืฉื™ืจ ื›ื•ืœืœ ืžืกืš ืžื’ืข ืงื˜ืŸ, ืžื™ืงืจื•ืคื•ื ื™ื, ื—ื™ื‘ื•ืจ 5G ื•ื—ื™ื™ืฉืŸ ื˜ื‘ื™ืขืช ืืฆื‘ืข. ืœื•ื—ืฆื™ื, ืžื“ื‘ืจื™ื ืขื ื”ืกื•ื›ืŸ, ื•ืขืœ ื”ืžืกืš ืžื•ืคื™ืขื” ื“ืžื•ืช ืื ื™ืžื˜ื™ื‘ื™ืช ืฉืžื™ื™ืฆื’ืช ืื•ืชื•. ืžืขื‘ืจ ืœืฉืืœื•ืช ื•ืชืฉื•ื‘ื•ืช, ื”ืจืขื™ื•ืŸ ื”ื•ื ืฉ-Muse ื™ื•ื›ืœ ื’ื ืœื‘ืฆืข ืคืขื•ืœื•ืช ื›ืžื• ืœื ื”ืœ ืžืฉื™ืžื•ืช, ืœืชื›ื ืŸ ื“ื‘ืจื™ื ื•ืœืขื–ื•ืจ ื‘ืงื ื™ื•ืช ื•ื‘ื”ื–ืžื ื•ืช. ืžืืจืง ืžืกื‘ื™ืจ ืฉื”ื ืœื ืœื’ืžืจื™ ืžื•ื›ื ื™ื, ืื‘ืœย ืžืชื›ื ื ื™ื ืœื”ืชื—ื™ืœ ืœืฉื•ื•ืง ืื•ืชื• ืœืงืจืืช ื“ืฆืžื‘ืจ. ื–ื” ื”ื™ื” ืฆืจื™ืš ืœื”ื™ื•ืช ื”ืจื’ืข ืฉืœ OpenAI ืœื ืžื˜ื.
ืžื˜ื ื—ื•ืฉืคืช ืžืฉืงืคื™ VR ื—ื“ืฉื™ื ื‘ืžืฉืงืœ 100 ื’ืจื ื‘ืœื‘ื“ - ื•ื™ืฉ ืขื•ื“ (ืคืœื•ืก ืขื“ื›ื•ืŸ ื—ืฉื•ื‘ ๐Ÿ‡ฎ๐Ÿ‡ฑ ืขืœ ื™ืฉืจืืœ!) ื‘ื›ื ืก Meta Connect 2026 ืžื˜ื ื”ืฆื™ื’ื” ืืช Meta VR Glasses, ืžืฉืงืคื™ VR ืงืœื™ื ืžืื“, ื›ืฉื”ืžืขื‘ื“ ื•ื”ืกื•ืœืœื” ื ืžืฆืื™ื ื‘ื™ื—ื™ื“ืช ืžื—ืฉื•ื‘ ื ืคืจื“ืช ื‘ื›ื™ืก.ย  ื”ื ื›ื•ืœืœื™ื ืžืกื›ื™ micro-OLED ื‘ืจื–ื•ืœื•ืฆื™ื™ืช 5K, ืฉืœื™ื˜ื” ื‘ืืžืฆืขื•ืช ื”ืขื™ื ื™ื™ื ื•ื”ื™ื“ื™ื™ื. ื”ืžื—ื™ืจ: 1,299 ื“ื•ืœืจ. * ื”ืขืจื”: ืœื ื™ื•ื“ืข ืื ื–ื• ืชื”ื™ื” ื”ืฆืœื—ื”, ืื‘ืœ ื–ื” ื‘ื“ื™ื•ืง ืžื” ืฉืืคืœ ื”ื™ื• ืฆืจื™ื›ื™ื ืœื”ืฆื™ื’. ืคืกืคื•ืก ืฉืœ ืืคืœ. ื‘ืžืงื‘ื™ืœ ื”ื ื”ื›ืจื™ื–ื• ืขืœย Ray-Ban Meta Audioย ืžืฉืงืคื™ื™ื ืœืœื ืžืฆืœืžื” (*ืฉื•ื‘ ื”ืขืจื”: ืื•ืœื™ ื›ื”ืžืฉืš ื™ืฉื™ืจ ืœื‘ื™ืงื•ืจืช ืขืœ ืคื’ื™ืขื” ืงืฉื” ื‘ืคืจื˜ื™ื•ืช), ื•ื“ื•ืจ ื—ื“ืฉ ืฉืœ Ray-Ban Meta ืขื ืฆื™ืœื•ื 3K ื•ืกื•ืœืœื” ืžืฉื•ืคืจืช. ืกื•ื›ืŸ ื”-AI ืฉืœ ืžื˜ื Muse ืžืงื‘ืœ ืฉื™ื—ื•ืช ืงื•ืœื™ื•ืช, ื‘ื™ืฆื•ืข ืžืฉื™ืžื•ืช ื‘ืจืงืข, ืื™ื ื˜ื’ืจืฆื™ื•ืช ืขื ืฉื™ืจื•ืชื™ ืฆื“ ื’ืณ - ื•ื’ื ื™ื›ื•ืœืช ืœืขื‘ื•ื“ ื“ืจืš ื”ืžืฉืงืคื™ื™ื. ืžื˜ื ื’ื ื”ืฆื™ื’ื” ืืช Muse Charmย - ืžื›ืฉื™ืจ AI ืงื˜ืŸ ืœื›ื™ืก ืฉืžืืคืฉืจ ืœื“ื‘ืจ ื™ืฉื™ืจื•ืช ืขื ื”ืกื•ื›ืŸ ื‘ืœื™ ืœื”ื•ืฆื™ื ืืช ื”ื˜ืœืคื•ืŸ. (* ืฉื•ื‘ ื”ืขืจื”: ื–ื” ืžื” ืฉ OpenAI ื™ืฆื™ื’ื• ื‘ืงืจื•ื‘. ืžื˜ื ื”ื–ื“ืจื–ื” ื•ืขืงืคื” ืื•ืชื) ื•ื™ืฉ ื’ื ื”ืงืฉืจ ื—ืฉื•ื‘ ๐Ÿ‡ฎ๐Ÿ‡ฑ ืœื™ืฉืจืืœ: ืžืฉืงืคื™ Ray-Ban Meta ื•-Oakley Meta ื™ื’ื™ืขื• ืจืฉืžื™ืช ืœื™ืฉืจืืœ - ืขื•ื“ ื‘ 2026. ื–ื” ืขื•ื“ ืจื’ืข. ืœืื•ืจ ืขื ื™ื™ืŸ ื”ืคืจื˜ื™ื•ืช - ื”ืจืฉื•ื™ื•ืช ื—ื™ื™ื‘ื•ืช ืœื”ื›ื™ืจ ืืช ื–ื”.
15
2
3
93
31,475
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
You know Wix Engineering on LinkedIn. Now come meet us on Instagram ๐Ÿ‘‹ Follow Wix Engineering for the code, the people, the lessons learned, and everything engineering: instagram.com/wix_eng/
2
3
385
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
AI context engineering, 756K+ production messages, 6B+ daily media requests, zero-downtime database migrations, infrastructure migrations across 1,000+ MySQL servers, our upcoming meetup and more! Hereโ€™s whatโ€™s inside our latest @Wix Engineering newsletter - ๐Ÿง  From Autocomplete to Intelligence: How Language Became the Operating System of AI / @GiladBarkan AI can write code, use tools, and power autonomous agents - but underneath it all, it still predicts the next word. Explore how language became the operating system of AI, why Transformers and scale changed everything, and where todayโ€™s models still fall short. ๐Ÿšฆ Donโ€™t Offer What Canโ€™t Be Done: Deterministic Executability Gating for LLM Skill Selection at Scale / @AshkenaziOrtal, Vitalii Kloz & Mykhailo Ulianchenko Semantic matching alone isnโ€™t enough in production. Across 756K+ messages, Wixโ€™s executability gate filtered out 59.4% of matched skills, reduced skill context size by 90.5%, and saved 228.8M tokens by checking what users can actually do before skills reach the LLM. Now as a new paper on @arXiv. ๐Ÿ› ๏ธ [Upcoming Meetup] Migrations at Scale: Code, Infrastructure & Automation / @Danaavital58 & Shmuel Mekonen Two migrations, two very different challenges: automating code changes across hundreds of codebases, and migrating 1,000+ MySQL servers across 160+ clusters in 30 days - with zero downtime. [Oct 28, at Wix Campus TLV] ๐Ÿ–ผ๏ธ Architecting for 6 Billion Daily Requests: Inside Wixโ€™s Media Platform / Tomer Salton What happens when image resizing needs to work across 300M+ websites and more than 6 billion media requests every day? Go behind the architecture powering real-time AVIF encoding, smarter caching, lower resize costs, and secure asset delivery at massive scale. ๐ŸŽฏ Attention is Expensive: Optimizing Context Windows for Precision and Cost / Maya Halevy More context doesnโ€™t always mean better AI. Learn how attention limits, token waste, and the โ€œLost in the Middleโ€ problem impact LLM accuracy - and how better context design can improve reliability while reducing cost. ๐Ÿ—„๏ธ How We Built a Zero-Downtime Database Migration Service at Wix / Bar Shauli How do you move critical production databases between clusters without users noticing? Explore the Kafka, ProxySQL, and MySQL architecture behind Wixโ€™s internal migration service - from streaming data to cutting over live traffic with zero downtime. Read it here ๐Ÿ‘‡ wix.engineering/so/2dQ3DoCRl
3
2
386
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
Really impressive work from @Wix data science team, who published research based on a real challenge they saw and solved in our own customer care agent. They researched how to help AI agents distinguish between skills that are relevant to a userโ€™s question and those that can actually be executed for that specific user. This helps users get more relevant and useful support from the start. The team was able to find a more precise approach that reduced the information the model needed to process by 90.5% and saved 228.8 million tokens. A huge achievement. Congratulations to @AshkenaziOrtal, Vitalii Kloz, Mykhailo Ulianchenko, and the entire team.
[new @arxiv paper] Thereโ€™s a huge gap between playing with AI skills in a demo and โš™๏ธ engineering them for production. In a simple setup, semantic routing feels like enough: the user asks something, the LLM matches the text to a skill description, and executes it. But in real-world production at scale (like @Wix Customer Care), semantic matching alone isn't enough. ๐Ÿšฉ Why? Because semantic matching only looks at what the user says, completely ignoring what the user is actually allowed or able to do. A user asking to "connect a domain" is a 100% semantic match for the Domain Skill. But if they don't have an active premium plan, executing that skill is a dead end. In a new arXiv paper by @AshkenaziOrtal, Vitalii Kloz, and Mykhailo Ulianchenko, we present Helpmateโ€™s (Wixโ€™s Customer Care AI agent) architecture: introducing a deterministic executability gate that evaluates account state before candidate skills reach the LLM. Tested across 756K+ user messages in production: ๐ŸŽฏ 59.4% of semantically matched candidates were filtered out because the account state made them non-executable ๐Ÿ“‰ 90.5% reduction in skill context size 228.8M tokens saved ๐Ÿ›ก๏ธ In a targeted replay of 1,000 risky conversations, removing the gate caused the model to select a non-executable skill in 7.8% of cases. Read the full paper: arxiv.org/html/2608.01050v1
3
1
1
18
2,922
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
[new @arxiv paper] Thereโ€™s a huge gap between playing with AI skills in a demo and โš™๏ธ engineering them for production. In a simple setup, semantic routing feels like enough: the user asks something, the LLM matches the text to a skill description, and executes it. But in real-world production at scale (like @Wix Customer Care), semantic matching alone isn't enough. ๐Ÿšฉ Why? Because semantic matching only looks at what the user says, completely ignoring what the user is actually allowed or able to do. A user asking to "connect a domain" is a 100% semantic match for the Domain Skill. But if they don't have an active premium plan, executing that skill is a dead end. In a new arXiv paper by @AshkenaziOrtal, Vitalii Kloz, and Mykhailo Ulianchenko, we present Helpmateโ€™s (Wixโ€™s Customer Care AI agent) architecture: introducing a deterministic executability gate that evaluates account state before candidate skills reach the LLM. Tested across 756K+ user messages in production: ๐ŸŽฏ 59.4% of semantically matched candidates were filtered out because the account state made them non-executable ๐Ÿ“‰ 90.5% reduction in skill context size 228.8M tokens saved ๐Ÿ›ก๏ธ In a targeted replay of 1,000 risky conversations, removing the gate caused the model to select a non-executable skill in 7.8% of cases. Read the full paper: arxiv.org/html/2608.01050v1
2
5
2
10
3,779
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
Thank you @HebrewU_heb for highlighting Shmuel Mekonen's work, article and upcoming meetup: 1 - @HebrewU_heb - alumni.huji.ac.il/event/wix_โ€ฆ 2 - Article: 1,000 Servers. 160 Clusters. 30 Days. Zero Downtime - Migrating @Wix's MySQL Fleet to Graviton wix.engineering/post/1-000-sโ€ฆ 3 - Meetup: Migrations at Scale: Code, Infrastructure & Automation wix.engineering/migrations-aโ€ฆ
[Meetup] Large-scale migrations are never just about moving code or infrastructure. They're also about doing it without ๐Ÿšฉ slowing everyone else down. At @Wix, two very different migration challenges required two very different approaches: 1 - Rolling out automated code changes across hundreds of codebases, while keeping developers focused on day-to-day work. 2 - Moving 1,000+ MySQL servers across 160+ clusters, multiple regions, and two major platform transitions - all within 30 days and with zero downtime. Dana Avital and Shmuel Mekonen will share how these migrations were designed and executed: the architecture behind them, the automation that made them possible, the tradeoffs, the failures along the way, and the key decisions that helped them succeed. Join us on October 28 at 17:30 at Wix Campus TLV: wix.engineering/migrations-aโ€ฆ
1
2
431
Morad Stern ๐Ÿ‡ฎ๐Ÿ‡ฑ retweeted
ืื ื™ ืื•ื”ื‘ ืœืขื‘ื•ื“ ืขืœ ื”ืžืง ื•ืœืฉื™ื ๐Ÿค˜ ืžื•ื–ื™ืงื”, ืื‘ืœ ืชืžื™ื“ ื”ืคืจื™ืข ืœื™ ื”ื ืชืง ื”ื–ื”: ื”ืžื•ื–ื™ืงื” ื™ื›ื•ืœื” ืœื”ืฉืชื•ืœืœ, ืื‘ืœ ื”ืžืกืš ื™ื™ืฉืืจ ืื“ื™ืฉ ืœื’ืžืจื™. ืื– ื‘ื ื™ืชื™ ืืคืœื™ืงืฆื™ื™ืช macOS ื—ื“ืฉื” ืฉืžืื–ื™ื ื” ืœืกืื•ื ื“ ืฉืžืชื ื’ืŸ ื‘ืžื—ืฉื‘, ืžืืคืœ ืžื™ื•ื–ื™ืง ื•ืกืคื•ื˜ื™ืคื™ื™ ื•ืขื“ ื™ื•ื˜ื™ื•ื‘ ื‘ื“ืคื“ืคืŸ - ื•ื”ื•ืคื›ืช ืื•ืชื• ืœืืคืงื˜ื™ื ื—ื™ื™ื โœจ ืฉื‘ื•ืงืขื™ื ืžืงืฆื•ื•ืช ื”ืžืกืš. ืฆื‘ืขื™ื, ื—ืœืงื™ืงื™ื ื•ืชื ื•ืขื” ืฉืžื’ื™ื‘ื™ื ื‘ื–ืžืŸ ืืžืช ืœืžื•ื–ื™ืงื” ื•ืœืงืฆื‘, ืžื‘ืœื™ ืœื”ืกืชื™ืจ ืืช ื”ืชื•ื›ืŸ ืฉืื ื™ ืขื•ื‘ื“ ืขืœื™ื•. ืคืชืื•ื ื”ืžื—ืฉื‘ ืฉืœื™ ืœื ืจืง ืžืฉืžื™ืข ืืช ื”ืžื•ื–ื™ืงื”, ื”ื•ื ืดืžืจื’ื™ืฉืด ืื•ืชื” ืื™ืชื™! ื”ืืคืœื™ืงืฆื™ื” ืชืขืœื” ื‘ืงืจื•ื‘ ืœ-App Store. ืื ืชืจืฆื• ืฉืืขื“ื›ืŸ ืืชื›ื ื›ืฉื”ื™ื ื‘ืื•ื•ื™ืจ ืฉืœื—ื• ืœื™ ื”ื•ื“ืขื”.
6
2
1
53
6,569
ื‘ืคืื ืœ ื‘ื‘ื™ืช ืกืคืจ, ื›ืฉืจื•ื‘ ืžื™ ืฉืฉื ืงื˜ื™ื ื™ื, ื”ืฉืจ ืฉืื—ืจืื™ ืขืœ ืžืขืจื›ืช ืื›ื™ืคืช ื”ื—ื•ืง ืžื—ืœืง ืžื—ื–ื™ืง ืžืคืชื—ื•ืช ืฉื”ื•ื ืคื•ืชื—ืŸ ืœืžืฉืงืื•ืช ืขื ืืœื›ื•ื”ื•ืœ ื•ื‘ืฆื•ืจืช ืืงื“ื—. ื›ื”ื ื™ื–ื ืขื‘ืจื™ื™ื ื™ - ืงื•ื•ื™ื ืœื“ืžื•ืชื•. ืื‘ืœ ื”ืืฉืžื”? ืจืง ืขืœ ืžื™ ืฉืžื™ื ื” ืืช ื”ืื“ื ื”ืžืกื•ื›ืŸ ื”ื–ื”.
87
150
9
2,030
17,023