HNHacker News
TopNewBestAskShowJobs

pbronez

2,117 karma · joined May 10, 2019

[ my public key: https://keybase.io/peterbronez; my proof: https://keybase.io/peterbronez/sigs/aK3o2l59iuXf4ZE61-arwlc3CsPE1jYuB6pu_vjfi0c ]

<a rel="me" href="https://hachyderm.io/@PeterBronez">Mastodon</a>

submissionscomments
pbronez··on Creepy Crawlies
“Expect to lose some functionality, at least when accessing our resources anonymously.”

This seems fine to me. It would be a better world if we could have anonymous bulk data access. But if aggressive scrapers are bloating host costs, I’m fine with logging in.

Now, the flip side is that ONCE logged in, I want my bulk access. The worst of all worlds with when you demand authentication and then STILL block bulk access.

Case in point, I want to automatically download my Amazon and Target order records. This is easy to automate with playwright or whatever, but authentication stays annoying. My sessions expire quickly and I have to re-auth all the time. There should be an API to pull this data down.

pbronez··on My Business Is Dying
Looks interesting. Might want to double check your website - the Customization / Assistants / MCP pills in your hero aren't working for me
pbronez··on The turbulent AI era is here
Did you publish your results anywhere? I’d be interested to read about them in detail.
pbronez··on Nebula Sans
This surprised me, so I looked it up. Basically YouTube pays ~55% and Nebula pays 50%.

This is complicated by Nebula's "creator owned" situation. Which is complicated [0]. It's something like:

  - Nebula's assets live in Watch Nebula LLC

  - Watch Nebula LLC is owned by 83% Standard Broadcast and 16% Curiosity Stream (CURI)

  - Standard Broadcast "is owned by 44 total people, all of whom are creators"
So it's fair to say there are creators with significant equity stakes in Nebula. It's NOT fair to assume that the 50% of revenue left after share is distributed evenly among all creators on the platform via an equity mechanism.

Of course then there's Floatplane, which is entirely owned by LMG. At least a couple years ago they targeted "a 70/30 split" but arrive there via "a fixed cost menu for features [...] based on bandwidth/maintenance costs' [1]

I'm glad YT is still sharing a meaningful amount of money with creators.

[0] https://medium.com/@cameron-paul/who-actually-owns-nebula-95...

[1] https://linustechtips.com/topic/1335443-floatplane-creator-s...

pbronez··on WebMCP: Teaching Your Website to Talk to AI Agents
I'd be interested to read a blog post with more details about how this works.
pbronez··on Ask HN: Why do corporate failures always seem to punish the wrong people?
Yeah no layoffs in that example. The connection is that it's very easy for the carpenters to see the Etsy business as "crucial," but the CEO decided to kill that entire line of business. From an individual career perspective, you could have mis-identified this and focused on being the guy who manages the Etsy storefront.

Selling the Etsy side of things is an interesting idea. I wonder how you could do that without also selling off the tools and things you need for the SKUs you want to keep. I don't think they had distinct brands, for example.

pbronez··on How much of HN is AI?
Invite sent, be kind & have fun
pbronez··on How much of HN is AI?
Sure do - invite sent to the email indicated by your HN profile.
pbronez··on How much of HN is AI?
I got you - invite sent to the email on your HN profile.
pbronez··on Don't Wordle
It does - you can type them but it rejects the word when you submit it.

I think this DOES make it way harder than regular wordle.

pbronez··on Ask HN: Why do corporate failures always seem to punish the wrong people?
The trick is to align with how corporate leadership defines "crucial" and not the way a "crank turner" defines it.

A crank turner thinks having a solid platform to stand on, a long crank arm and good lubrication is crucial. Because those things ARE crucial to crank turning.

Corporate management is looking at the machine the crank is attached to and thinking if that machine is going to produce the company's next quarter of growth.

The machine you crank could even be very profitable right now, but it's not where the company is going. Here's an easy to follow example from a smallish business:

https://youtu.be/9EMR2BVbG8Q

pbronez··on Hister – A private, full content search index that you control
I've been using LinkDing + SingleFile for this. Nice to have another option!
pbronez··on I Just Want to Search
If I quote a search terms, every returned result must contain that string. Even if that means showing zero results. Anything else is madness.
pbronez··on Being ambitious and being a dad
Yeah that biological clock had a surprisingly loud alarm!
pbronez··on AI is removing the middle class of software engineering?
Good talk, thanks for sharing! I’m trying something similar for personal project I’m hacking on.

Specifically, I’m spending a lot of time on the first few instances of a pattern that I hope to have the AI scale out for me. My thought is that I can provide an opinionated project structure, feel it out myself, then point the AI to those working examples in the future. Until then, I’m mostly just using ai as fancy autocomplete + domain research.

pbronez··on Mojo 1.0
Aren’t they already separated? Max is an inference runtime, written in Mojolang, which uses the MLIR compiler chain to target diverse accelerator hardware.
pbronez··on Grok Bot
Interesting that the pricing is "Cursor Ultra" and "Cursor Premium Teams." This very much feels like the Code --> Cowork product iteration.
pbronez··on Protopia
WWZ the movie takes almost nothing from the book beyond the title. Huge missed opportunity. The book uses a future history format and digs into practicalities. For example, people invent a new hand weapon optimized for killing zombies. They need it because firearms are too expensive, require too much skill to employ effectively in civilian areas. Still, conventional arms prove essential for large scale progress. IIRC much of the book focuses on the struggle to get from the initial chaos to manageable battle lines.

Overall I really enjoyed it. I should re-read it…

pbronez··on DeepMind's WeatherNext model achieves breakthrough forecasting cyclones
Cool how they integrated both huge machine-scale data and smaller human-curated data for this project.

> The model was co-trained on two distinct data modalities: global weather dynamics and expert-curated historical cyclone observations. By training end-to-end on nearly 20 terabytes of global atmospheric data and the historical IBTrACS database spanning nearly 5,000 historical storms, the model learns complex atmospheric patterns and how to model extreme weather.

pbronez··on DeepMind's WeatherNext model achieves breakthrough forecasting cyclones
Insightful paper, thanks for sharing. Two things stand out to me.

First, it reinforces that you want methods that get better with more data. It emphasizes that the current approach cannot improve based on historic data - that’s the opportunity that ML based approaches exploit.

Second, it highlights that mature legacy solutions are tough competitors. They benefit from extensive tuning and real world feedback. Even when you have a genuinely better approach, it will take meaningful time & effort to achieve the current standard.

Patience and solid long term strategy are needed to make progress in these situations. You need confidence that your approach will win long term, backed by enough money & time to prove yourself correct.

pbronez··on Launch HN: ProvenMetal (YC S26) delivers circuit boards in days instead of weeks
It’s tricky to compete with your own suppliers. For example, see Microsoft’s dance with Dell & HP when they launched Surface. How do you plan to keep the CMs happy while you’re operating your own line?
pbronez··on Launch HN: ProvenMetal (YC S26) delivers circuit boards in days instead of weeks
Some financial engineering could help here too. That abandoned inventory is useful to somebody, but warehousing can be non trivial… how do you get liquid and reroute the parts?

There’s some flavor of consignment waiting to happen here.

pbronez··on Launch HN: ProvenMetal (YC S26) delivers circuit boards in days instead of weeks
Credit based on a SBIR is a great idea. Seems like something the Small Business Administration should offer.
pbronez··on I regret migrating to Codeberg
Because most forges sell private repos as their primary business model.

I encounter Codeberg as a forge first, then have to dig in to discover that it's a highly opinionated service with a different business model than most similar entities.

pbronez··on Show HN: Run an 80B Qwen in 4.3 GB of RAM on a Mac, and a 35B on an iPhone
There are many search APIs available, I like Kagi's.

Microsoft and Amazon both provide web snapshot services that purport to give you a kind of agent-first internet archive. You can approximate something like that using common crawl, but it's a huge amount of data. Downloading the internet is impossible or a bad idea for almost everyone.

pbronez··on There Will Come Soft Rains (1950) [pdf]
Yeesh how bleak
pbronez··on Don't be a meat proxy
And this is fine! This is the whole point of living in a civilization. Master your craft, get useful, trade with people who know other useful things, everyone is better off.

You could just as easily point out that subsistence farmers exist. That’s great, but we still need industrial agriculture to feed everyone. No single human fully understands that domain either… even ignoring the chemistry, biology and meteorology, those farms use digitally controlled tools! That means all of computer science is embedded in farming.

It’s great to pursue broad knowledge. It’s important to understand your dependencies and respect the abstractions you build on. It’s pure hubris to think an individual auteur can challenge a multi-billion person civilization.

pbronez··on SQLite Critical CVEs or LLM Slop?
I once had a customer threaten to reject a delivery over a CVE. I dug into it. The CVE was that an Ocaml PostgreSQL client didn’t implement an authentication feature.

I convinced the customer to accept the delivery by pointing out that (1) our app had zero lines of ocaml and (2) the feature had been implemented in the ocaml driver since the CVE was issued.

pbronez··on About the security content of macOS Tahoe 26.6
Pretty extreme solution… you can get Claude models from AWS Bedrock and Google Model Zoo. These are both very helpful for compliance and security, but do require you to have a cloud strategy.
pbronez··on A $500 RL fine-tune of a 9B open model beat frontier models on catalog review
Yes. Fine tuning is an optimization. Like all optimizations, it’s downstream of figuring out your problem, implementing a naive solution, scaling the naive solution, and getting frustrated by SWAP-C. THEN you start poking at what optimizations are possible, choosing an approach, implementing the optimization, and redeploying.

This cycle happens when you have a well defined use case with high volume. It is the far opposite end of the spectrum from the general purpose intelligence on tap that frontier AI models purport to deliver.

I think focused fine tunes and big general models will coexist. Ideally with smart routing and caching to use small, specialized and local options when appropriate.

← PreviousPage 2 of 34Next →