Technophile, strategist, entrepreneur, racer, and dad.

Marin County, CA
Joined December 2010
Peter Guagenti retweeted
A highlight from NVIDIA GTC Berlin: Jensen Huang spotlighting our work with @baseten on post-training specialized models for code review. Proud to work with NVIDIA and Baseten to push the frontier of AI for software development!
2
7
4
29
1,663
Peter Guagenti retweeted
PHP users: patch today. Two critical RCEs (CVSS 9.2) affect "all PHP installations serving web requests". Update to 8.5.12, 8.4.27, 8.3.36 or 8.2.35.
We have posted more information about today's security releases, including the sequence of events that led to the discovery, how the PHP security team is improving their processes, and what we need from you to continue this critical work. #php #security thephp.foundation/blog/2026/…
2
7
36
2,527
Peter Guagenti retweeted
AI coding tools keep getting better. What are we getting better at? We sat with @kunalstwt to talk about the skills software engineers should invest in: understanding systems, solving useful problems, and taking responsibility for what ships. Catch the full 1hr 16 video below!
5
7
2
25
8,736
Peter Guagenti retweeted
New home for your code, same rabbit on review. CodeRabbit is a launch partner for Cursor Origin's app marketplace, bringing AI code review directly to Origin-hosted pull requests. coderabbit.ai/blog/coderabbi…
Origin, Cursor’s code hosting platform, now supports over 20 integrations. Connect your repos to tools for CI, deployments, code review, security, monitoring, and project management.
4
23
6
42
4,829
Peter Guagenti retweeted
i'm open-sourcing the code behind the coderabbit music video. although the template was laid out by @donaldjewkes and @other__reality's videos, there are a few things that were done differently in this video to obtain these results: 1. word-level lyric timing. we split the vocals out with Demucs and force-aligned the lyrics with stable-ts. P(doom)'s karaoke estimates word timing from line length. the alignment also found a sung "as mentioned above" at 2:00 that wasn't in the lyric sheet. a custom f-aligner was buitl 2. we swapped the watercolor (p5.brush) for a spider-verse print look: halftone shading, color-plate misregistration for depth, ink that boils 12x a second, characters on twos. all 3,088 frames render in ~5 min on a laptop. P(doom)'s guide budgets 2.5 s per frame. this is a style reference change more or less 3. no karaoke bar. each line picks a mode which is a huge stacked type for the hook, comic caption boxes for the rabbit's inner voice, ransom-note cutouts at 3am, or typed into the scene ("@coderabbitai pause" in a comment box). x autoplays muted, so it has to read with the sound off and to also get a catchy hook 4. did a LOT of research. one agent audited the AI-twitter zeitgeist (the Navier-Stokes blowup vortex became the rabbit hole). another studied spider-verse technique and K-pop direction, which is why hop/nope/yep get the exact same framing every chorus. a total of 8 video subagents and 3 research agents ran in cursor to achieve this 5. like P(doom), parallel agents built the chapters (8 here), working from a written style sheet and a storyboard locked to the measured beats. the difference was the review loop in which the main agent pulled contact sheets from the encoded mp4 and sent weak shots back with notes 6. just under 4 hours from the raw mp3 to the final encode, in one Cursor session with Claude Opus 5.5. STYLE_SHEET, STORYBOARD and ANIMATION_GUIDE are in the repo if you want to point it at your own song. my total opus costs from my cursor dashboard stand around $200 here's the github repo: github.com/arimanyus/hophopn…
one-shotted this music video using opus 5.5 (life of coderabbit's rabbit)
5
3
2
120
9,531
Peter Guagenti retweeted
Know what your PR does. CodeRabbit Change Stack - your explainability layer: > Overview + Layers: map the PR's logical changes > Findings: issues, proposed fixes and agent prompts > Code Peek + AI Chat: definitions, context, answers > Blast radius: understand the impact
3
8
3
25
3,870
Peter Guagenti retweeted
CodeRabbit is available in Claude Marketplace! Apply a portion of your committed Anthropic spend to CodeRabbit, the same route we used for Vercel Sandbox and Workflows! Read Anthropic's announcement: claude.com/blog/how-coderabb…
You can now discover tools, agents and expert partners on Claude Marketplace. Use it to: - Add connectors and plugins like Slack and Notion - Buy agents and products from companies like Cursor and CrowdStrike - Scale with service partners like Accenture and Deloitte
2
4
2
15
3,058
Peter Guagenti retweeted
Beyond the benchmarks, what was Opus 5.5 like to work with? Hendrik and Gowtham discuss an overnight coding run, the token tradeoffs, and why higher reasoning effort didn’t consistently catch more bugs.
We ran Opus 5.5 through CodeRabbit's review pipeline. > On 80 known bug patterns it caught 51 vs 49 for our production mix. > On 13 harder cases, 10 vs 5. > It found a retry-count race in Cal​.com that production missed. The catch is ~50% more tokens, and 9 bugs our baseline caught that Opus 5.5 didn't. Read our full Opus 5.5 model evaluation here: coderabbit.ai/blog/opus-5-5-…
1
4
2
22
3,367
AI has solved how software gets written, but it also broke how software is *shipped*. With 1,000's of PRs issued by agents, how do you trust the code that has been generated? @coderabbitai solved this with a single platform that reviews, explains, triages, and secures every PR:
Agents can create PRs in minutes, but understanding them still takes work. CodeRabbit Change Stack connects a PR's purpose, behavior, dependencies, and code so reviewers can really verify it before approving. Do you understand what you're about to merge?
29
Peter Guagenti retweeted
We ran Opus 5.5 through CodeRabbit's review pipeline. > On 80 known bug patterns it caught 51 vs 49 for our production mix. > On 13 harder cases, 10 vs 5. > It found a retry-count race in Cal​.com that production missed. The catch is ~50% more tokens, and 9 bugs our baseline caught that Opus 5.5 didn't. Read our full Opus 5.5 model evaluation here: coderabbit.ai/blog/opus-5-5-…
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
10
15
9
185
48,037
Peter Guagenti retweeted
We’re honored to be named a 2026 AI40 winner! The IA40 recognizes the 40 most important private companies in applied AI. Thanks to @Madrona, @AWSstartups, @Microsoft, @GoogleStartups, @NYSE, @Delta, & @McKinsey for the recognition!
The 2026 Intelligent Applications 40 is here. Now in its sixth year, the IA40 recognizes the top private companies shaping the future of applied AI, from foundational infrastructure to full-stack intelligent applications. This year's list reflects a market maturing fast: → 450+ companies nominated and voted on by 72 investors across 54 leading firms → $410B raised by this year's 45 winners since founding — more than 25x what the inaugural 2021 class had raised at the same point → 23 repeat winners and 22 new entrants, a 49% turnover rate, down from 68% last year, as some of 2025's breakout companies start to entrench The big pattern this year is how companies are “Harnessing the Value of AI.” As frontier models become more capable and commoditized, the winners are increasingly the companies proving they can turn AI capability into real ROI, whether they're building the application, aggregating the models, or supplying the infrastructure underneath. Databricks remains the only company to appear on every IA40 list since we launched it in 2021. Congratulations to all the 2026 IA40 winners. Your work is defining what's possible in this new era of intelligent applications. Explore all 45 winners and our analysis of the companies, categories, and funding patterns defining this year’s list → ordnl.link/RAFgnA7 Thank you to our great partners: @AWSstartups, @GoogleStartups, @Microsoft, @NYSE, @Delta, and @McKinsey.
4
6
1
22
3,361
Peter Guagenti retweeted
😂🧑🏻‍💻
65
880
89
6,555
477,159
Just wrapped up this super enlightening episode of the MLOps Community podcast featuring Peter Guagenti, a total tech guru who's really shaping the AI scene in software development.
1
2
1
5
1,302
A8: My biggest concerns with AI remain privacy, respect for copyright, and monopoly concentration of power. Big tech has already proven they will exploit whatever they can get away with to maximize profits. I worry about the negative impact on small co's and creators. #eweekchat
1
128
A7 (cont'd): As for AI to avoid, IMO ChatGPT is overused for use cases it's not great at. Go find the built-for-purpose tool that solves your problem best! #eweekchat
101
A7: Shameless plug for our own AI coding assistant, @tabnine. As someone who is very rusty but does still touch code occasionally, it's a life saver. Not GenAI, but I'm a big fan of @madkudu for sales & marketing. @Adobe is also doing great things. #eweekchat
85
A6: I would expect te chips to head towards commoditization and for the costs to drop dramatically. This will unlock more innovation and the integration of AI into more applications and work processes. I would expect affordable AI at the edge to emerge relatively soon #eweekchat
66
A5: My advice to tech pros trying AI tools is to come with an open mind, try it all, and be prepared to adapt how you work. Resistance to change is holding many people back, and like every other wave of innovation risks leaving behind the people who won't learn. #eweekchat
66
A4: I have used AI in my work for years; models to identify ICP, to optimize messaging, & prioritize sales & mktg activities. With GenAI, we added list building, email outreach generation, meeting summarization, and image generation. All have huge impact on outcomes. #eweekchat
61