HNHacker News
TopNewBestAskShowJobs

jpadkins

1,532 karma · joined December 22, 2010

jesse dot adkins at gmail posts reflect my personal opinion only, not my employer.
submissionscomments
jpadkins··on The Lost Treasure of Sid Meier's Pirates
I played Pirates! as recently as 2022. Definitely not lost treasure for me.
jpadkins··on Cursor launches Origin, GitHub alternative
I don't follow EU politics. Isn't that position a pretty common EU political view? What is evil about decentralization of power?
jpadkins··on Memory prices climb 500% in 12 months
200 pages https://www.goodreads.com/en/book/show/3028.Economics_in_One... There are free PDFs if you don't want to buy it.
jpadkins··on Cursor launches Origin, GitHub alternative
why?
jpadkins··on GLM-5.3: Frontier coding with emergent cyber capabilities
Piggybacking on this thread to ask my question: What are alternatives that are multiplayer (team oriented) by default? For example, I want my team to see all my sessions easily, vise versa. another way of stating: all the agents are running in a container that that any member of the team can view and interact with.
jpadkins··on Gemini 3.7 Flash
welcome to the internet, not all browsers are the same
jpadkins··on Gemini 3.7 Flash
It's not an assumption that the customer is Google scale that led the $35k monthly price point. It was that you (and many others like you) are breaking the terms of service (which state you are not supposed to store, process or analyze the search results). It was an API built for a different era, that doesn't really exist anymore.

> It has honestly been easier to use Anthropic to help me build my own search index & crawling infrastructure than deal with Google.

That's great to hear! FWIW I agree with you that it's harder for independent professionals to get started on Google dev services than others, but the Search API is not developer service and was never intended to be.

jpadkins··on Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs
So it hasn't innovated since Oct of 2000? (when AdWords launched)
jpadkins··on Advancing the price-performance frontier with GPT‑5.6
When model intelligence reliably hits 90%-95% of current day knowledge worker tasks, they are going to burn those weight to silicon and we will see another 10X improvement in price/performance frontier.

The dynamic GPU clusters will be used for the 5% of tasks, and pushing out the frontier. Also there will be a set of knowledge tasks that are not done today (because they are too difficult for most knowledge workers), that will start being done in the future.

jpadkins··on LearnVector – Andrew Ng's AI company building one‑to‑one learning experiences
I had the same thought. Then I thought, maybe "replace" is what will become opposite. Antonyms of "replace" are hold keep precede break damage deprive forfeit lose penalize remain take worsen

Hmmm. Maybe they should rethink their motto.

jpadkins··on Advertise in ChatGPT
I think there are rules about public disclosures in a run up to an IPO. They may not be able to publish updates until the official IPO process begins.
jpadkins··on Codex is wearing out our devices
This reddit thread is devoid of any useful information. Just fear and speculation.
jpadkins··on Show HN: Jacquard, a programming language for AI-written, human-reviewed code
it pales in comparison to all the brain cells burned chasing dumb ideas. Don't sweat it, and keep burning tokens in an effort to discover something great.
jpadkins··on You can't unit test for taste
I think another important question is can you distill taste? (another comment uses the phrase "externalize", which might mean something similar).

I think people have been trying for the written word, with some degree of success (anti-slop skills). I have been trying for visuals, and it's pretty meh. It's easy to get a multimodal LLM to follow a style guide, but a style guide doesn't capture everything that accounts for taste. And anything that is dynamic (not a screenshot test) seems really hard or really expensive.

jpadkins··on There are a few things that I look back on as my mistakes in the early days
also this post was a reply to Sandy on X.
jpadkins··on The founder's playbook: Building an AI-native startup
Hah. Poe's law.
jpadkins··on SpaceX to buy Cursor for $60B
Why not N+1 planets? Too many extinction level possibilities outside of our control.
jpadkins··on SpaceX to buy Cursor for $60B
money is just an idea that spreads (a meme).
jpadkins··on SpaceX to buy Cursor for $60B
What you are describing applies to all forms of money. It has value because people believe other people will use it as money. If that belief drops, the value drops.

People comment on gold and Bitcoin, but don't realize the same principles apply to US dollars and bonds.

jpadkins··on Stanford grads walk out on Google CEO Sundar Pichai speech
Brin did not come out of retirement or steer the ship :-)

He probably came out of retirement because LLMs were showing promise to a life long dream.

jpadkins··on Stanford grads walk out on Google CEO Sundar Pichai speech
okay, but that is not the world that you live in. Many countries are religious and use religion to decide immigration.
jpadkins··on I'm Eric Ries, author of "The Lean Startup" and new book "Incorruptible" – AMA
I don't have any questions for you. Just wanted to thank you for writing Lean Startup, it was very influential to me.
jpadkins··on When AI Builds Itself: Our progress toward recursive self-improvement
why only a few thousand? serious question.
jpadkins··on Gemma 4 12B: A unified, encoder-free multimodal model
local models still need information retrieval.
jpadkins··on Claude Opus 4.8
I find this site useful https://artificialanalysis.ai/leaderboards/models
jpadkins··on Google employee charged with $1M Polymarket insider trading bet on search term
That is the beauty of the Resolution part of prediction markets. If evil DDT Corp bets money to skew the market, then they lose even more money on the Resolution (assuming the resolution is deterministic of harm and has not been manipulated).
jpadkins··on Citing 'severe' math deficits, UC faculty demand a return to SAT tests for STEM
who uses SAT scores as "potential succeed"??
jpadkins··on Google employee charged with $1M Polymarket insider trading bet on search term
Nice try to make a law sound moral.
jpadkins··on Google employee charged with $1M Polymarket insider trading bet on search term
The other question you could be asking: Why does Google want to keep this secret for a period of time?
jpadkins··on Google employee charged with $1M Polymarket insider trading bet on search term
yes it is good in many scenarios.

Imagine bad (incorrect and potentially harmful) information is public knowledge. Examples are "X cures cancer" or "Is Y dangerous to consume".

A prediction market will be seeded by public knowledge (of course it cures cancer or its safe to consume), which you describe "suckers". History is filled with many examples of bad public knowledge that turned out to be false (e.g. DDT is safe pesticide).

An insider (someone who knows the drug trial results, or works at the Corp creating the harmful substance) is incentivized to trade on that knowledge, which creates a better informed public (via people who pay attention to prediction markets).

Why does secret(insider) knowledge exist? To the benefit of the organization that wants to keep the knowledge secret. Insider trading laws purpose is to keep Corp and gov orgs in power. They prevent the dissemination of true information (for private power). Prediction markets incentivize the dissemination of true information, a public good.

← PreviousPage 2 of 29Next →