Hacker News: dalberto

New comment by dalberto in "Cloudflare Quick Tunnels"

dalberto — Fri, 18 Sep 2026 20:52:04 +0000

I built something for this exact use-case using Cloudflare Tunnels:

I also use CF managed auth to make auth easier.

New comment by dalberto in "pg_durable: Microsoft open sources in-database durable execution"

dalberto — Fri, 05 Jun 2026 20:48:14 +0000

> Postgres as it's built now lacks any semblence of extensibility

PostGIS. pgvector. TimescaleDB. Citus. pg_cron. pgmq. Apache AGE. ParadeDB. hstore. plv8. postgres_fdw. pg_partman. pg_stat_statements...

The extension API is the thing making your thesis possible. Rewriting it away would mean deleting the exact feature you're asking for.

New comment by dalberto in "Show HN: A new benchmark for testing LLMs for deterministic outputs"

dalberto — Wed, 29 Apr 2026 17:55:28 +0000

A benchmark without Opus 4.6/4.7 feels incomplete.

Show HN: TakeHome – LLC vs. S-Corp tax calculator for solopreneurs

dalberto — Mon, 16 Mar 2026 17:04:15 +0000

I got tired of doing rough napkin math on whether to elect S-Corp for my LLC, so I built a calculator over the weekend.

Most tools I found either ignore state/city taxes, charge you for the answer, or require creating an account to see anything useful. TakeHome runs entirely in your browser (no sign-up, no analytics, no server-side storage) and updates every number in real time as you drag sliders.

What it models: - LLC self-employment tax vs. S-Corp FICA on W-2 salary

- QBI deduction (Section 199A) with SSTB phase-out

- Solo 401(k) with SECURE 2.0 rules (traditional/Roth/split, catch-up limits by age bracket)

- Self-employed health insurance deduction

- S-Corp admin/compliance cost comparison

- NY State income tax, franchise tax, PTET

- NYC UBT (LLC), GCT (S-Corp), PIT with IT-219 credit

You can save scenarios, compare any two side by side (it shows exactly which inputs differ and the dollar impact), and run multi-dimensional "what-if" experiments.

There's also an AI mode where you describe a scenario in plain English and it generates the experiment for you.

The tax engine is ~2,000 lines of TypeScript with no external dependencies. Stack is React 19, MobX, Tailwind v4, deployed on Cloudflare Workers. AI feature uses Claude Haiku.

Caveats: NY/NYC taxes are fully modeled; for other jurisdictions it's directionally correct at the federal level (SE vs. FICA, QBI, 401k analysis are location-independent). Standard deduction assumed. SSTB assumed for QBI. Not financial advice, treat it as prep for a conversation with your CPA.

I've added detailed tooltips on every computed line item with formulas, IRS code references, and source links. Every number shows its work.

Interested in feedback, especially if you spot tax logic errors. Also curious whether I should prioritize other states (CA, TX, FL?) or itemized deduction support next.

https://takehome.money

Comments URL: https://news.ycombinator.com/item?id=47401674

Points: 2

# Comments: 1

New comment by dalberto in "Postgres extension complements pgvector for performance and scale"

dalberto — Tue, 30 Dec 2025 17:19:00 +0000

I generally agree that one database instance is ideal, but there are other reasons why Postgres everywhere is advantageous, even across multiple instances:

- Expertise: it's just SQL for the most part - Ecosystem: same ORM, same connection pooler - Portability: all major clouds have managed Postgres

I'd gladly take multiple Postgres instances even if I lose cross-database joins.

Vibe Coding's Last Mile Problem

dalberto — Mon, 10 Nov 2025 20:38:06 +0000

Article URL: https://blendingbits.io/p/vibe-codings-last-mile-problem

Comments URL: https://news.ycombinator.com/item?id=45880623

Points: 1

# Comments: 0

New comment by dalberto in "Why we migrated from Python to Node.js"

dalberto — Mon, 03 Nov 2025 19:04:22 +0000

Async django is not for the faint of heart, but it's definitely possible in 2025.

I recently wrote about issues debugging this stack[1], but now I feel very comfortable operating async-first.

[1] https://blendingbits.io/p/i-used-claude-code-to-debug-a-nigh...

I Used Claude Code to Debug a Nightmare

dalberto — Sat, 01 Nov 2025 17:37:55 +0000

Article URL: https://blendingbits.io/p/i-used-claude-code-to-debug-a-nightmare

Comments URL: https://news.ycombinator.com/item?id=45783578

Points: 3

# Comments: 1

New comment by dalberto in "Leaving serverless led to performance improvement and a simplified architecture"

dalberto — Fri, 17 Oct 2025 16:38:15 +0000

Haven't deployed production Java in years, so I won't speak to it. However, even with Go's static binaries, I'd like to leverage the same build and deploy process as other stacks. With Docker a Go service is no different than a Python service. With Docker, I use the same build tool, instrument health checks similarly, etc.

Standardization is major. Every major cloud has one (and often several) container orchestration services, so standardization naturally leads to portability. No lock-in. From my local to the cloud.

New comment by dalberto in "Leaving serverless led to performance improvement and a simplified architecture"

dalberto — Fri, 17 Oct 2025 16:32:59 +0000

Even when running things in their own box, I likely want to isolate things from one another.

For example, different Python apps using different Python versions. venvs are nice but incomplete; you may end up using libraries with system dependencies.

New comment by dalberto in "Leaving serverless led to performance improvement and a simplified architecture"

dalberto — Wed, 15 Oct 2025 20:39:01 +0000

Hard disagree. I've used Docker predominantly in monoliths, and it has served me well. Before that I used VMs (via Vagrant). Docker certainly makes microservices more tenable because of the lower overhead, but the core tenets of reproducibility and isolation are useful regardless of architecture.

New comment by dalberto in "GPT-5: Key characteristics, pricing and system card"

dalberto — Thu, 07 Aug 2025 19:25:53 +0000

Consider: https://github.com/musistudio/claude-code-router

or even: https://github.com/sst/opencode

Not affiliated with either one of these, but they look promising.

What I Learned from Building and Shutting Down an AI Legal Startup

dalberto — Thu, 27 Feb 2025 14:54:56 +0000

Article URL: https://blendingbits.io/p/beyond-the-fine-print

Comments URL: https://news.ycombinator.com/item?id=43194882

Points: 3

# Comments: 0

New comment by dalberto in "Launch HN: Hatchet (YC W24) – Open-source task queue, now with a cloud version"

dalberto — Thu, 27 Jun 2024 17:01:11 +0000

That makes sense, though a bit disappointing. One hope of using Postgres as a task queue is simplifying your overall stack. Having to host RabbitMQ partially defeats that. I'll stay tuned for the Postgres-backed messaging!

New comment by dalberto in "Launch HN: Hatchet (YC W24) – Open-source task queue, now with a cloud version"

dalberto — Thu, 27 Jun 2024 15:20:38 +0000

I'm super interested in a Postgres-only task queue, but I'm still unclear from your post whether the only broker dependency is PostgreSQL. You mention working towards getting rid of the RabbitMQ dependency but the existence of RabbitMQ in your stack is dissonant with the statement 'a conviction that PostgreSQL is the right choice for a task queue'. In my mind, if you are using Postgres as a queue, I'm not sure why you'd also have RabbitMQ.

New comment by dalberto in "Show HN: Hatchet – Open-source distributed task queue"

dalberto — Fri, 08 Mar 2024 19:01:13 +0000

I'm curious if this supports coroutines at tasks in Python. It's especially useful for genAI, and legacy queues (namely Celery) are lacking in this regard.

It would help to see a mapping of Celery to Hatchet as examples. The current examples require you to understand (and buy into) Hatchet's model, but that's hard to do without understanding how it compares to existing solutions.

New comment by dalberto in "Show HN: Struct – A Feed-Centric Chat Platform"

dalberto — Fri, 01 Mar 2024 01:26:30 +0000

Looks very cool, planning on giving it a try. Also appreciate transparency around pricing.

New comment by dalberto in "Wear OS "Hybrid" design has two OSes, two CPUs, "100 hour" battery life"

dalberto — Wed, 28 Feb 2024 16:05:47 +0000

That's fair, a lot of these features do not require a beefy processor. I think the value lies in the AW being a full-featured platform. Its processor is overkill for the lightest use-cases like remote shutter, but for others it's nice having a smartphone-like experience.

Case in point: Maps are full featured, it's not just about traffic, it's also about points of interest, supporting different modalities (crucially, public transit), and generally feeling like the full maps experience.

Like any other platform, 3rd party apps are nice for niches that the manufacturers do not serve, or do not serve well. For example, lifting apps make it super easy to log sets, reps, rest, etc; useful data the core experience does not offer.

As mentioned earlier, I think there's room for both. I'm actually not opposed to sporting a Garmin for endurance sports, where battery life is king :)

New comment by dalberto in "Wear OS "Hybrid" design has two OSes, two CPUs, "100 hour" battery life"

dalberto — Wed, 28 Feb 2024 15:46:35 +0000

Some of the features I use on my Apple Watch all the time:

- Turn-by-turn directions with full-featured map apps.

- Stream and control podcasts.

- Ditto Audiobooks.

- As a remote shutter for my phone camera (eg group selfies).

- Dedicated 3rd party fitness apps (Alltrails, strength training apps).

I understand all this functionality is not everyone's cup of tea, but there's clearly a market for it.

Show HN: Privacy Scout – understand privacy policies in minutes with LLMs

dalberto — Wed, 15 Nov 2023 15:11:36 +0000

At Atticus AI, we're on a mission to make legal understanding more accessible to everyone. After building our contract analysis tool and talking to folks, we noticed a common use-case: understanding privacy policies. We decided to build a standalone tool, Privacy Scout, to address this need.

With Privacy Scout, all it takes is a URL to a privacy policy. From there, the tool takes over by:

- Retrieving and analyzing the policy.

- Grading the policy with an easy-to-understand and transparent scoring system.

- Allowing users to compare privacy policies across companies in our public directory.

Big learnings for us:

- Updating our privacy policy was necessary due to the lack of key terms.

- A significant portion of the work in effectively utilizing LLMs involves data engineering, such as data munging and developing robust pipelines.

- Achieving standout performance with Retrieval Augmented Generation (RAG) requires extensive domain-specific experimentation.

We're continuing to build a suite of tools to make legal understanding more accessible, and I would love your feedback!

Comments URL: https://news.ycombinator.com/item?id=38277453

Points: 8

# Comments: 1