Baz gives development teams a clear understanding of how every change aligns with their specs, designs, standards and protocols.
External link for Baz
Boston, Massachusetts, US
Tel Aviv, IL
Hamburg, DE
Baz reposted this
חברת Baz, המפתחת כלי AI לבדיקת קוד באופן אוטומטי עם הצעת תיקונים. דורגה במקום הראשון בעולם בדיוק בדיקות קוד שנכתב על ידי AI. וקדימה את מעבדות הבינה המלאכותית המובילות OpenAI, Anthropic, Google, Cursor (קישור לכתבה בתגובות) Guy Eisenkot | Nimrod Kor | Ronny Hassid
Baz reposted this
Baz ranks first in precision in the new independent Code Review Bench index, outperforming tools from major AI labs and signaling a shift in how real-world developer performance is measured *** The results placed Baz ahead of tools developed by several of the world’s leading AI labs, including OpenAI, Anthropic, Google and Cursor. Baz also ranked second in the overall composite score, which combines precision and recall. https://lnkd.in/dw-8bDi9 Guy Eisenkot Nimrod Kor
Baz reposted this
Wow. Despite 25 years in this industry, I don't think I've ever had the privilege of being part of an initial 3rd party benchmark.... until today. Finding ourselves outpacing industry giants like GitHub and Cursor is a remarkable, and validating, moment here at Baz. More than that, it's a testament to the collaborative work being with our customers as we tackle #AICodeReview together. We're proud to rank: #1 in precision 🎯 #2 in F1_Score ⚖️ #1 in F_0.5_Score 🤹 Thank you to Shriyash U., Etan Ginsberg, and the entire team at Martian for publishing this critical industry standard. #AgenticCoding
Baz reposted this
This is probably the most exciting moment since Baz was founded. Our review precision is ranked above Anthropic, OpenAI, GitHub, and any other tool in the market in the first independent AI code review benchmark. Unlike things you might have seen in the past, this is not a vendor comparison or an internal evaluation. It is the first independent benchmark dedicated to AI code review agents. Precision here is measured by the percentage of comments developers actually act on. It is a real-world proxy for signal over noise. That has always been our lighthouse. From day one, we believed code review should not be about generating more comments. It should be about generating the right ones. As a Product Manager, seeing that thesis acknowledged publicly is incredibly meaningful. Thank you Martian, for putting us in the spotlight 💜 We are just getting started 🚀 Nimrod Kor Guy Eisenkot #AgenticCoding #CodeReview #AI #DevTools
We just had a room full of PMs and designers on the 38th floor in Tel Aviv, overlooking the sea at sunset talking about technical Product Management in 2026. If there was a takeaway, it’s this: product development is an information system. Once you view it that way, it becomes clear that AI’s primary impact is not speed alone. Everyone has a new role to play. Huge thanks to Tali and Shaked for going deep, and to Cato Networks for hosting 🍕🍻🍺 and to our event partners: Shachar Azriel, Zvi Fried, Netanel Sananes, Maayan Brosh
This isn't going to be for everyone! The Autonomous Codebase Maturity Roadmap is our forward-thinking framework for introducing autonomy into GitHub without compromising trust or security. When you pull back the curtain on GitHub's automation configurations you discover there's plenty of small tweaks you can already consider if you feel your team is ready for the next level of SDLC automation. Configure responsibly and tell what you think.
The number one requirement for running Auto-Fix isn’t write access. It’s precision 🎯 If code reviews are verbose, poetic, noisy, vague, or wrong, you cannot act on it in your repository. Fixer, our new coding agent for fixing bugs in PRs is already tightening security-sensitive paths, normalizing identity boundaries, reconciling schema contracts, enforcing permission logic and resolving subtle UI state edge cases. There's still plenty of work but the path forward is much clearer now and its precision 🎯 first. Enable Fixer now: https://baz.co/agents
Routine pull request fixes are one of the biggest drains on engineering time: Failing tests. Lint noise. Minor refactors... Developers end up acting as human patch machines. We built the Fixer Agent to eliminate that layer of friction. It handles routine PR corrections automatically. It validates the fix against tests and review feedback. It keeps changes minimal and traceable. It does not guess and hope. It fixes and proves. This is part of a bigger direction we are building toward: an autonomous codebase. Read more here 👇 https://lnkd.in/dhgchF8r
Baz reposted this
כל PM מכירה את הרגע הזה: כתבת דרישות. הפיתוח מימש. מורג׳ג׳ ודופלט. ואז את פותחת את הסביבה, ואומרת לעצמך - “אה… אבל איפה ה...״, ״זה לא בדיוק...״ ומשם מתחיל הסבב: תגובות, הבהרות, סלאק, עוד review, עוד תיקון… והפיצ’ר מתעכב. בתקופה האחרונה אני מתעסקת הרבה בשאלה איך AI יכול לעזור דווקא בנקודה הזו - לא לכתוב קוד במקום מפתחים, אלא לחבר בין ה-intent של ה-PM למה שבאמת נבנה ויוצא לפרודקשן. על זה בדיוק אדבר במיטאפ: From Spec to Shipping. אם אתם אנשי מוצר שחשוב להם שדברים יהיו על הוואן, נראה לי שתתחברו 🙂
Join us for an upcoming 🏖️ Tel Aviv meetup "AI for PMs: From Specing to Shipping" – where we decode how AI agents orchestrate features end-to-end, spot design-to-implementation drifts, and turn every PM into a feature shipping machine without the ping-pong drama. 🗓️ When: Tuesday, February 24 5PM 📍 Where: Cato Networks TLV HQ, Floor 38, Tel Aviv (epic views included) 🎤 Speakers: Tali Garsiel (Staff Engineer @ Cato Networks) and Shaked Edri (PM @ Baz) sharing product hacks. RSVP here: https://luma.com/bdliftol Zvi Fried Netanel Sananes Maayan Brosh Shachar Azriel Nimrod Kor
Fixer is now live: A powerful coding agent for fixing code review issues. By configuring the correct sandbox environment you give Fixer the exact context your code requires downstream. Fixes are composed, tested and executed inside that environment so changes are validated against the same standards that gate your merges. Your setup is what turns Fixer into a power tool for preventing code review rework. Shaked Edri, Product Manager at Baz, explains how to get a Fixer running in seconds 👇👇👇