Hacker News: CaveTech

New comment by CaveTech in "I built a vulnerable app and spent $1,500 seeing if LLMs could hack it"

CaveTech — Thu, 04 Jun 2026 19:54:14 +0000

The harness does not matter that much, it's getting leaner every cycle.

New comment by CaveTech in "I built a vulnerable app and spent $1,500 seeing if LLMs could hack it"

CaveTech — Thu, 04 Jun 2026 19:53:11 +0000

You've confused what I wrote, we are in agreement. The fact codex found the vulns means that mythos almost certainly will.

New comment by CaveTech in "I built a vulnerable app and spent $1,500 seeing if LLMs could hack it"

CaveTech — Thu, 04 Jun 2026 02:52:25 +0000

I’m highly confident that prior exposure is irrelevant at this point. I work on vulnerability detection at a hyperscaler.

New comment by CaveTech in "I built a vulnerable app and spent $1,500 seeing if LLMs could hack it"

CaveTech — Thu, 04 Jun 2026 01:59:12 +0000

It was found with gpt 5.5 7/10 times it’ll be trivially found by mythos

New comment by CaveTech in "What twenty years of DevOps has failed to do"

CaveTech — Sun, 18 Jan 2026 04:02:03 +0000

These things are not nearly equivalent. It’s writing code, it’s not software engineering.

New comment by CaveTech in "Why can't transformers learn multiplication?"

CaveTech — Fri, 24 Oct 2025 23:48:39 +0000

I never made a claim that it's ineffective, just that it's of limited effectiveness. The diminishing returns kick in quickly, and it's not applicable in more domains than it is applicable.

New comment by CaveTech in "Why can't transformers learn multiplication?"

CaveTech — Fri, 24 Oct 2025 22:49:03 +0000

The latter, and I would disagree that “this works and scales well” in the general sense. It clearly has very finite bounds by the fact we haven’t achieved agi by running an llm in a loop..

The approach of “try a few more things before stopping” is a great strategy akin to taking a few more stabs at RNG. It’s not the same as saying keep trying until you get there - you won’t.

New comment by CaveTech in "Why can't transformers learn multiplication?"

CaveTech — Fri, 24 Oct 2025 20:38:38 +0000

The amount of paths in the wrong direction are infinitely more than then number in the right direction. You'll quickly realize this doesn't actually scale.

New comment by CaveTech in "Show HN: Autism Simulator"

CaveTech — Thu, 02 Oct 2025 05:59:14 +0000

[citation needed]

New comment by CaveTech in "Windsurf employee #2: I was given a payout of only 1% what my shares where worth"

CaveTech — Sat, 26 Jul 2025 02:33:29 +0000

As you describe this is largely a cash flow business and the bulk of the value should be extracted via dividends to the benefit of major shareholders.

A tech enabled business needs gross margins north of 70% to be attractive from a leverage standpoint, unless revenue is scaling very rapidly. Without these there’s no attractive exit opportunities.

New comment by CaveTech in "Windsurf employee #2: I was given a payout of only 1% what my shares where worth"

CaveTech — Sat, 26 Jul 2025 01:00:18 +0000

Sorry to break it to you but 10% profit on 6.5M rev is very low and will absolutely not fetch a high multiple, especially considering this is a mature 10 year old business. This is not a high growth business and you may have grown overly rose colored glasses by thinking it could be priced as one.

New comment by CaveTech in "Travelers to the U.S. must pay a new $250 'visa integrity fee' – what to know"

CaveTech — Fri, 18 Jul 2025 15:20:06 +0000

There are border towns where some workers do daily crossings. Without refund infrastructure in place it looks like this would add $250/day fee for all of these individuals

New comment by CaveTech in "Why email startups fail"

CaveTech — Tue, 01 Jul 2025 15:21:27 +0000

The whole thesis OP is making is that this isn’t really true, evidenced by real world behaviours. Electron apps have some of the highest market share, even the worlds most popular ide is an electron app.

The amount of people who won’t adopt based on pricipal is exceedingly small.

New comment by CaveTech in "Why email startups fail"

CaveTech — Tue, 01 Jul 2025 05:48:02 +0000

It is mentioned, there's an entire section named "The HEY Experiment".

New comment by CaveTech in "Why email startups fail"

CaveTech — Tue, 01 Jul 2025 05:45:30 +0000

The other part of the average is the 200m+ monthly active users who can't seem to find the uninstall button.

New comment by CaveTech in "Load Test GlassFlow for ClickHouse: Real-Time Dedup at Scale"

CaveTech — Sun, 22 Jun 2025 17:53:35 +0000

I think your benchmark may miss the mark a bit if this is your angle.

20m records and 9k/sec isn’t very impressive. I would imagine most prospective customers have larger workloads, as you could throw this behind Postgres and call it a day. FWIW I was interested but your metrics made me second guess and wonder what was wrong.

New comment by CaveTech in "Base44 sells to Wix for $80M cash"

CaveTech — Thu, 19 Jun 2025 17:05:24 +0000

More than the team with no product. There's a huge difference between large entrenched businesses and early-stage startups. Turnover in large corporations is huge, the team has turned over multiple times already, they can do so again.

An early stage company that barely has any code? Yes, in that case, it's the team that matters because they haven't fulfilled the vision yet.

New comment by CaveTech in "I'm Peter Roberts, immigration attorney, who does work for YC and startups. AMA"

CaveTech — Fri, 16 May 2025 18:29:11 +0000

Are there any difficulties in obtaining TN Visas related to the stage of the sponsor company?

More specifically, is it feasible for a Canadian operating startup to open a US entity and employ Canadian engineers under the TN Visa category? This would be an early stage non-VC backed company, but profitable and paying above prevailing market wage.

New comment by CaveTech in "Show HN: Hydra (YC W22) – Serverless Analytics on Postgres"

CaveTech — Sat, 10 May 2025 03:53:10 +0000

Current user of Timescale for events processing, with heavy use of materialized views for rolling aggregates.

Is this a use case that you think Hydra would be competetive on?

New comment by CaveTech in "Dbt Labs acquires SDF Labs"

CaveTech — Wed, 15 Jan 2025 04:55:00 +0000

It compiles to SQL, so in since sense yes it is schema dependent. As mentioned in the article DBT itself has no understanding of semantics, so it’s not enforced by DBT directly but if it doesn’t compile to working SQL expressions it’s obviously not going to work.