Hacker Newsnew | past | comments | ask | show | jobs | submit | mritchie712's commentslogin

Looks cool! It's a lot of work to get a full data stack set up and people are losing interest in stitching the pieces (ETL, warehouse, BI) together.

> Agents made bad inferences because they had no context on the business

We've been working on this since before the chatgpt launch.

We started with a semantic layer since there were already good open source options and LLMs at the time were good at writing the JSON (remember function calling?) to run a semantic query.

But as LLMs have gotten smarter and people wanted to do more data work in agents, we found we needed something more flexible, so we built an "Ontology" that lets you store all the terms you use in your company and connect them to the data points (e.g. tables, columns, metrics) that matter.

https://www.definite.app/blog/ontology-ai-analytics


For anyone that doesn't use twitter:

I index all my local Claude Code sessions in DuckDB. I have 202,381 messages in the last 30 days.

There's been a steady increase since Opus 4.6 in the model saying "honest".

It probably shouldn't, but this bugs me.

Should I assume most of the time you're lying and you're being honest in this one message?

I was pumped in the first few hours of Fable where this had seemingly been "fixed". 100+ messages and no "honest" to be seen. But it didn't last.

Within a few hours, Fable proved itself to be the most honest model to date.

Here is the rate at which visible assistant text contained the string "honest" (case-insensitive), split by model:

  claude-fable-5:             25 / 1,397   = 1.7895%
  claude-opus-4-8:            83 / 5,818   = 1.4266%
  claude-opus-4-7:           163 / 16,432  = 0.9920%
  claude-opus-4-6:            18 / 5,877   = 0.3063%
  claude-haiku-4-5-20251001:   0 / 71      = 0.0000%
  claude-sonnet-4-6:           0 / 4       = 0.0000%

I wonder how this would be looked upon by the ever changing rules of claude code.

If someone from Anthropic sees this, would love to know if I can use my max plan here.


all of our customers are using Anthropic APIs for programmatic use. Codex and other providers let you use Oauth. But inside of a sandbox, you can technically use max plan since it is the same as using Claude locally.


Can you provide some info on how to use the max plan / ChatGPT pro subscription via OAuth? I was about to subscribe but wasn't entirely sure how to configure that


> The day was to begin like any other, with Antigravity open

> This unexpected shift completely broke my preferred workflow

it might not have been so unexpected if you knew you were one of ~15 people that start their day with Antigravity


students get a couple free years of a Google One subscription which comes with access to the nicer models and some free claude credits to use inside Antigravity IDE. Though there's a small chance I'm wrong and everyone has access to those models.

The free tokens is the only reason why I've been using Antigravity. And after this god awful update (I had the exact same scenario as the article author) and once my free student plan runs out, I'm immediately switching to something else.

Makes me wish we could regulate these big tech companies and run them big fines whenever they abuse their monopoly by unnecessarily downgrading the experience for users. I hate how all their stocks continue to climb despite headline after headline of new updates breaking things for users. It's like they can just do whatever they want and never get punished for it meaningfully.


That was a surprising sentence. But, Antigravity is fine. I mean, I only open it because it's the only way Gemini is going to get tool use right on the first try, but it works like all the other VS Code forks (acceptable, not great). I don't mind using it, and if Google AI is your one AI subscription, then Antigravity is obviously the editor to use, since Gemini fails to play well with others.


> The day was to begin like any other, with Antigravity open (yes, there are tens of us!)

I think they know.


don't think that was there when I read it the first time


I've become an Antigravity convert mainly because of the generous limits, especially on the Anthropic models. User since day 1.


They used to have good limits that lasted hours, now I wiped mine in a couple of minutes..


for reference, Flipper Zero was $199.

does anyone know how much they're thinking for Flipper One?


https://gizmodo.com/the-company-behind-the-flipper-zero-hack...

>> Flipper’s goal is to sell the device for around $350.


With that money going to Russia.

Just sayin'.

EDIT: The above is a perfectly factual statement about the Flipper. The downvotes are presumably coming from the pro-Russia bots and shills.


The entirety of your comment has no basis in fact.

Speaking as someone who watched the first flipper prototypes being soldered by hand.


Flipper One founders live and operate from London, as is the company. Pavel Zhovner has (or had at least) Ukrainian citizenship.

Terrible to always see misinformation from people who don't even check basics.


Yes there are a lot of pro-Russia bots and shills with 1000+ karma on HN. FOH


Where would you prefer it to go?


Before or after the AI collapse of 2026/27. I would say at least $499 without the addition of inflated memory pricing.


grand at min


> Can I use DuckDB with Quack as the catalog database for DuckLake?

> Not yet, but we are working on it!

Seems like a niche use case, but it's the one I'm most interested in.

Our lakehouse uses ducklake with postgres as the catalog. Seems like a DuckDB / Quack catalog would be an excellent alternative.


Well, we are really working on it: https://github.com/duckdb/ducklake/pull/1151

So you'll be able to test it in a few days.


Does this mean I can finally connect to a ducklake instnace hosted remotely? i.e. DuckLake is writing to disk on the remote server and my client is just a client.

Because rn even with Postgres as a catalog my client needs access to the underlying storage to use Ducklake.


Yes, Quack resolves this problem. In particular, your client (likely a DuckDB instance) will talk to a remote DuckDB that both has access to the underlying storage and can also serve as the catalog itself.


already works now! just tried it out


I think that Quack will become the primary option for a DuckLake catalog in the future, for several reasons. To list a few:

1. No type mismatches for inlining. If you use a non-DuckDB catalog, many types do not have a 1:1 mapping, which introduces additional overhead when operating on those data types.

2. You get the raw performance of DuckDB analytics (and now transactions) over the catalog. DuckDB reading DuckDB is simply faster than any of our Postgres/SQLite scanners.

3. No round-trip for retries. We can easily(tm) run the full retry logic on the DuckDB server side. Right now, these retries trigger multiple round trips for Postgres, making it a performance bottleneck for high-contention workloads.

Disclaimer: I'm a duckdb/ducklake developer.


This. Type casting is an insidious problem (both correctness, and perf)


We previously had our (https://www.definite.app/) agent write reports / dashboards in a YAML spec that would get rendered by our frontend framework (i.e. charts and tables).

For example, user says "build a report with revenue and orders by month and show 100 most recent orders". The agent would write a spec that would get rendered by our frontend.

This runs fast, but we were drowning in feature requests for what the framework could render (e.g. "I don't want labels here", "I DO want labels there", "can this chart be a heatmap", etc.)

A few months ago, we let the agent just write HTML instead. It takes longer to generate, but you get unlimited customization.

There are a host of issues with the new approach (non-technical users debugging a monstrous app they created), but net-net our customers like it much better.


How do you protect against prompt injection in this case?


for reference, it's the 2nd fastest model tracked in the "Highlights" section of https://artificialanalysis.ai/


Yes, it's incredibly fast. Openrouter is clocking 60 tokens per second, which is on par with the likes of sonnet, opus, GPT 5.5.


That section misses Cerebras and Groq which are up to 5x faster.


Very different tech and limitations though so wouldn’t make sense to compare 1:1 I think


What are the limitations ?


Much smaller context


looks like you work at github.

I completely understand a "people who give a shit stick around" mentality if you work there, but you can't expect users who run a business on it to stick around if it's broken.


I don’t think they were trying to hide that - they said they’re a “hubber” at the top. Maybe not obvious, but not obfuscated.


Correct, sorry I thought this was pretty obvious but in retrospect maybe not.

I'm not encouraging Mitchell to stay, I'm saying that my version of his post is about _me_ staying to make a brighter future, and adding my context on why I still believe that.

And finally I closed with "I hope we win you back" to be extra clear about it!


The problem is that you decided the correct place for your “version of his post” was in the comments to his version of his post. This rhetorically implies you’re offering your version as an alternative to his, and it explicitly sets up your version as a comment on his version.

And then you claim your version is only about you. Why post it as a reply, if that’s true? It would be off topic. But it isn’t off topic; because actually you are talking about OP in addition to yourself.


I don't like the "I canceled my x subscription" hype posts, but I did cancel Figma today. We've barely used it in months and this was the nail in the coffin.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: