Hacker News: stavarotti

New comment by stavarotti in "GLM 5.2 vs. Opus"

stavarotti — Mon, 22 Jun 2026 13:48:11 +0000

These style of comparisons are decent at showing capability but they don't really show me what I truly want - a sounding board and implementer with senior engineer-level execution. When I look back at all the teams that I've been part of, the best outcomes came from white-boarding (sometimes in the metaphorical sense) with one or two people, at times arguing, then finally compromising on a plan. Instead of synthetic benchmarks that try to be objective, I wonder if there's a way test this, or maybe I'm opining on a way of working that will soon be gone?

New comment by stavarotti in "Replies to comments on my "LLMs are eroding my career" post"

stavarotti — Mon, 08 Jun 2026 12:50:01 +0000

> On novel work:

> Work that introduces new methods, highly creative ideas, or solutions that have not been used or experienced before. More generally, an approach that introduces an innovative strategy to solve a complex problem.

Something that I've been thinking about for the past year or so is coming to grips with the fact that the vast majority (anecdote) of software engineering work is not novel (and maybe that's okay). Few opportunities lend themselves to doing truly novel work. Other than infrastructure work and highly specialized software, pause and ask yourself when you last encountered software were you said "how the hell did they do that?" or "damn, that's nice" (for me, the most recent was Ghostty). I think much of the angst that people have when they fear for their job is coming to the realization that LLMs can do most of the "standard" work that a lot of highly compensated individuals currently do. We've built livelihoods around this and the threat of that coming to an end is genuinely frightening.

New comment by stavarotti in "'Project Hail Mary' Crosses $300M in Sales to Become Amazon/MGM's Highest-Gross"

stavarotti — Mon, 30 Mar 2026 01:19:26 +0000

The movie was great. I'm always fascinated by adaptations, especially what they choose to exclude. I thought the movie struck the right balance.

New comment by stavarotti in "100 Jumps"

stavarotti — Fri, 13 Mar 2026 11:15:10 +0000

Such a fun game. It has the right amount of whimsy. Those shifty eyes…

New comment by stavarotti in "Atlassian to cut roughly 1,600 jobs in pivot to AI"

stavarotti — Thu, 12 Mar 2026 02:48:02 +0000

We use Jira. We're trapped by the processes built around it. Once it's entrenched, it's rather difficult to remove. I eagerly await for something that can replace Jira, but fear that it too will be bastardized to fit the process for however long the process lives.

New comment by stavarotti in "Amazon holds engineering meeting following AI-related outages"

stavarotti — Tue, 10 Mar 2026 13:41:09 +0000

I'm curious whether this is due to their insistence on using home grown tools ie, Kiro and not Codex/Cursor/Claude et all? I tried Kiro and quickly left.

New comment by stavarotti in "Claude struggles to cope with ChatGPT exodus"

stavarotti — Sun, 08 Mar 2026 19:31:39 +0000

I've largely found codex and claude code to be about the same however, codex tends to "think" harder and for longer which depending on the task, yields better results without too much steering.

On an unrelated note, UI is such a personal preference that it's impossible, beyond core pillars that have been studied for decades, to say one is better over the other. That being said, I like OpenAI's design system much better than Anthropic. OpenAI products (cli and chat ui) "feel" nice and consumer focused whereas Anthropic's products feel utilitarian and "designed for business".

New comment by stavarotti in "Claude Code wiped our production database with a Terraform command"

stavarotti — Fri, 06 Mar 2026 18:55:24 +0000

If you found this post helpful, follow me for more content like this.

I publish a weekly newsletter where I share practical insights on data and AI.

It focuses on projects I'm working on + interesting tools and resources I've recently tried: https://alexeyondata.substack.com

It's hard to take the author seriously when this immediately follows the post. I can only conclude that this post was for the views not anything to learn from or be concerned about.

New comment by stavarotti in "LLMs are a 400-year-long confidence trick"

stavarotti — Wed, 14 Jan 2026 13:29:25 +0000

Most of what I've been reading on either side of the argument is reductive. It's possible to have a take based on one's perspective and experience but it's impossible (at this time) to generalize things more broadly. I think what most people feel is multi-faceted: efficiency expectations from "leaders", job change inevitability (perceived or real), economic impact (should things not go well), loss of identity (am I a programmer, engineer, manager of things?), and several others. The discussions on the multiplicative effect of LLMs are being framed as a false dichotomy when it's far more complicated and nuanced.

New comment by stavarotti in "'Life being stressful is not an illness' – GPs on mental health over-diagnosis"

stavarotti — Sat, 06 Dec 2025 14:38:50 +0000

Life is stressful (in some respects, overly so), but we’ve dealt with this for millennia by having a strong support system. Not to be reductive over the multitudes of problems people face today, but most can and should be solvable by having a good support system. Family and good friends with whom you can speak frankly can do wonders. It doesn’t solve the affordability or job problems, but having someone to talk to, someone that you can trust and has lived experience, can go a long way. I can’t tell you how many times I’ve solved “problems” or at least lessened their impact, by consulting family and friends.

New comment by stavarotti in "The Math Legend Who Just Left Academia–For an AI Startup Run by a 24-Year-Old"

stavarotti — Thu, 04 Dec 2025 19:29:30 +0000

I always find this characterization baffling. Why does it matter if it's a 24 year-old vs 30, 35, or 50? Many aspects of life that we hold near and dear were created by very "young" people.

New comment by stavarotti in "John Giannandrea to retire from Apple"

stavarotti — Tue, 02 Dec 2025 04:27:43 +0000

I just want whoever is next in line to make Siri perform reliably and consistently. Not 70% of the time of even 90% of the time, 100% of the time for the limited uses cases that most (perhaps just me?) people use it for:

- Call "person"

- Call "business" (please don't say "I don't see so and so in your contacts" and on a second try, work)

- Find "place" (while driving) - Define "phrase or word" (please don't say I found this on the web)

- Set a timer or alarm

- Check the messages (in a sane way)

- Set reminders (this one surprisingly works well)

- Use intents correctly (I just want to be able to say "play 99% invisible in Overcast")

It doesn't need to do all the fancy things they show-cased last year. It just needs to do the basics really well and build from there.

New comment by stavarotti in "Google Antigravity just deleted the contents of whole drive"

stavarotti — Mon, 01 Dec 2025 12:42:24 +0000

An underrated and oft understated rule is always have backups, and if you're paranoid enough, backups of backups (I use Time Machine and Backblaze). There should be absolutely no reason why deleting files should be a catastrophic issue for anyone in this space. Perhaps you lose a couple of hours restoring files, but the response to that should be "Let me try a different approach". Yes, it's caveat emptor and all, but these companies should be emphasizing backups. Hell, it can be shovelware for the uninitiated but at least users will be reminded.

New comment by stavarotti in "Leak confirms OpenAI is preparing ads on ChatGPT for public roll out"

stavarotti — Sun, 30 Nov 2025 01:21:26 +0000

What if they get ads right? Not low rate garbage, but hyper targeted ads that are actually useful.

New comment by stavarotti in "SQLite as an Application File Format"

stavarotti — Fri, 28 Nov 2025 12:17:22 +0000

Previously: https://news.ycombinator.com/item?id=23508923

New comment by stavarotti in "Jane Goodall has died"

stavarotti — Fri, 03 Oct 2025 01:38:51 +0000

Such a wonderful soul. Few people have left an impression on me these many years later. I hope she's done the same for many here.

New comment by stavarotti in "Carved stone mask – Pre-pottery, Neolithic B period"

stavarotti — Sun, 17 Aug 2025 22:23:38 +0000

Every time I see artifacts like these I can’t help but think whether we are producing artifacts that will be discovered 10000 years into the future, and still be in a robust state.

Tell HN: It feels weird and but oddly good

stavarotti — Sun, 20 Jul 2025 15:17:59 +0000

I got access to Claude Max Pro few days ago with so I could use Claude Code to migrate a relatively complex app to a modern framework. I've tried various AI flows (ide/editor plugins, copy-paste chat windows, the usual), but I’d never tried the full "open a terminal and let her rip" approach.

I pieced together a workflow from half-remembered blog posts (there are too many ways to coerce the machine), and to my surprise it mostly worked. Once I got going, the weird part is that I felt more like an editor than a developer. Sure, I was steering the conversation (half paying attention and focused on other important things like letting the dog out) but mostly I was just pressing enter, not furiously typing out rebuttals like I've seen in many posts.

The strange part is that I tell the thinking machine what I want, hope for the best, and most of the time, I get what I want. Because I don't get what I want every time, the whole thing feels like the model is quietly saying "I have altered the deal, pray I don’t alter it any further". And when I'm surprised with with what I didn't ask for, I laugh and say to myself "this deal is getting worse all the time" and keep going.

Despite this, I'm enjoying Claude Code and will likely continue the subscription beyond my pilot period.

Comments URL: https://news.ycombinator.com/item?id=44625988

Points: 2

# Comments: 0

New comment by stavarotti in "Poll: Is AI Hype a Bubble?"

stavarotti — Sat, 15 Jun 2024 04:27:54 +0000

The generative aspect might be slightly oversold, but I’m interested in the practical applications being targeted by DeepMind and similar labs. AlphaFold’s ability to predict protein structures for disease treatment is very exciting.

New comment by stavarotti in "Largest Autonomous Ride-Hail Territory in US Now Even Larger"

stavarotti — Thu, 06 Jun 2024 05:23:28 +0000

I had my first ride in a Waymo car today. Despite closely following the field, nothing prepared me for the sheer magic of it. It just worked. After my fourth ride in a single day, all I could think was "hot damn!." It's easy to forget that, for most people, this is still the stuff of science fiction. Does it have room to improve? Sure. But today's experience was fantastic. I'm eager to see testing in my medium-sized Midwestern city (with all the snowy goodness). Also, the roads in SF are wild. I used to think Omaha streets were chaotic, but every visit to SF reminds me how much I appreciate the space and sprawl back home.