Ying-Yi Liang

I’ve been testing a few AI coding assistants across team workflows. The most capable one so far looks like Claude Code. OAI Codex CLI can sometimes crack tougher problems, but it’s missing a lot of practical features that Claude Code’s CLI has. Despite a lot of noise about them degrading the tool, I think it’s very capable. Better use of tools, making regular commits, opening PRs with good descriptions, and actually sticking to the todo-based plan without drifting off into “on the one hand, on the other hand” territory. Good bot! Anthropic also added analytics last week, following Cursor, which has had it for about a month. But for team use you need a solution that works in IDE and terminal. Cursor does this, with analytics and a web-based agent too. With those pieces in place, Cursor feels like a more complete package you could roll out across a team, something most alternatives don’t offer yet. This matters because analytics gives visibility into how these tools are actually being used. Coinbase shared they adopted Cursor wholesale and have 40% ai code, and now they’re aiming for 60%. That kind of feedback loop is key if you want to make the case for rolling out across a full team. (https://lnkd.in/eHGvkAxC) That said, Cursor also rolled out Bugbot and auto-enabled it on one of my repos, flagged an issue in a PR, then pushed an unverified commit straight to main, which caused a merge conflict in the PR. All terrible practice. So while Cursor seems to have the most team-focused tool at the moment, I’d definitely avoid Bugbot for now. Curious how others are navigating all these options. Do you think teams benefit from one joined-up tool across everyone, or is it better to stay agnostic and let devs pick what works for them?

14

9 Comments

Huge thanks to Zhe Jia, Shuai Wang, Aditi Bamba, and Martin Wang from DoorDash for publishing this incredibly practical breakdown of their RAG implementation for Dasher support. This post saved me countless hours of trial and error. The two-tier guardrail approach is what really clicked for me: * Run a cheap semantic similarity check first (sliding window comparison between response and KB articles) * Only if that fails, escalate to the full LLM evaluator Result: 90% fewer hallucinations, 99% fewer compliance issues, without killing your latency budget When you're actually shipping RAG to users, this kind of practical filtering makes all the difference. Everyone talks about prompt engineering, but nobody talks enough about how to catch bad outputs before users see them. Their broader setup—quality monitoring across 5 dimensions, regression test suites, the whole LLM Judge framework—gives you a real blueprint beyond the basic tutorials. You can't fix what you're not measuring. The article is https://lnkd.in/dbC9BG7a

2

Indeed. A small number of big tech companies have grossly overinvested in LLMs and overextended themselves -- even laying off tens of thousands of critically needed human employees and enshittifying their primary products and services to reallocate investment into it -- shooting Nvidia's profits and shares to the moon -- all in an insane attempt to "be first" for something that has no forseeable compelling ROI. It's a collective delusion and hype bubble. This is markedly different from the strategic "operate in the red to build marketshare" strategy of the few survivors of the dot-com bust, because at least those survivors were able to articulate (and realize) a logical ROI after reaching a tipping point of critical mass. This is more like the 98% of dot-com startups that burned bright fueled by VC right until they burned themselves out because they never actually had any believable plan for monetizing anything they were building. The lack of ROI will come rudely crashing down on these companies over the next several quarters as they run out of tricks to cook the books and hide the losses. Shareholders will get spooked and back out, share prices will crash, and big tech companies that have bet the farm on LLMs will suddenly find themselves without the necessary employees, capital, quality products/services, or reputational sway needed to save themselves. And when they collapse, Nvidia will collapse with them. And then maybe, finally, we can get back to the business of making things of actual value to humans.

19

rqlite -- and the way I develop it -- is fundamentally changing. LLMs are in the loop -- from ideation to Pull Request. https://lnkd.in/ejwXGddR

27

1 Comment

Update: Google is going back to in-person onsites for SWE roles in the next 2-3 months. I published a piece a few weeks ago where I summarized AI's effect on eng interview processes after surveying ~50 FAANG interviewers on interviewing.io, and in that piece I said there was no indication that any of the FAANGs were going back to in-person. (Piece linked in the comments.) Then I got this update from one of our interviewers last week. Things change fast. I'll share them as I learn more!

450

45 Comments

Big news, Slang AI has raised a $36M Series B, led by USVP. I stepped into the engineering manager role around this time last year and helped the team push through some tough quarters, tight deadlines, and shifting priorities to get here. This engineering team is technically strong, ships fast, and genuinely has each other's backs. They never stopped delivering, and they make me proud every day. Restaurants handle a massive volume of calls every day, but nearly half go unanswered. That's missed reservations, missed private events, and missed opportunities to connect with guests before they even walk in the door. Slang AI acts as the Superhost for every restaurant we work with, answering calls 24/7, booking reservations, handling questions, and capturing event leads with the kind of warmth and professionalism guests actually expect. Proud to be part of a team building technology that strengthens hospitality at scale. slang.ai/b #RestaurantTech #VoiceAI

86

3 Comments