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

You can now let ChatGPT interact with any service that exposes an API, and then additionally provides an MCP server for to interact with the API


For me it's being able to fully browse movies and tv shows, along with their universal rating and not some vague "You'll like it" nonsense.


My biggest issue with "cheaper" alternatives is the same pathway they all take.

Start cheap, gather market, then crank the costs after lock-in.

Even "open-source" is abused. First everything is open-source, and then reasons come out for why premium services will be closed source.


This one isn't quite saying its cheaper, or even charging, I think you might get a laugh if you click through. I don't think we'll need to worry about the costs being cranked after lock-in.


Is the word baas so horrified to people that they won't even click on the link to see that the code is a MASSIVE 1k loc (just using irony hehe)

But in all seriousness, I may be going on a tangent but I don't think that anybody can monetize code under less than 1k loc. Are there any cool examples anybody want to share?

Maybe "simple" api's would generally be the only thing that would be more monetizable and still fall under less than 1k loc. But still I would love hearing more about this kind of thing.


There are smart contracts in the Ethereum and Binance network making millions a day extracting transaction fees with much less than 1k loc. The code is even public.


Woah. Didn't expect it.

Can you give me some examples. Also I personally feel as if most of these are saturated and I don't think that I could earn a million with such loc. and maybe its me but personally I don't like touching most crypto since its grift. And the only one I'd like is monero for privacy but I doubt how much I need it anyway.


>Start cheap, gather market, then crank the costs after lock-in.

"cost"/"price" vocabulary clarification, should you ever want to read or write business plans, communicate with accountants, CFO's, etc.

"costs" are what companies pay for supplies/inputs that the company purchases.

"prices" are what those same companies offer to charge buyers for the products the company sells.

companies want to keep costs down, and companies benefit from high prices. (when you said "crank the costs", it thunks)

since people don't generally operate their lives as companies, it tends to seem like "costs" and "prices" are the same thing, but in addition to the above, "costs" to a company reflect actual expenditures in total, and "prices" represent an advertisement for each of something pending that has not transacted yet.

"cost" is an accounting term, total revenues - total costs = total profits

"price" is a marketing term, $1 each, $10 for a dozen!

(of course this could be quibbled into incomprehensively, which is another thing you should not do in "business communication", always streamline communication to get to the takeaway as quickly as possible)


What's with the dropped benchmark performance compared to the original o3 release? It was disappointing to not see o4-mini on it as well


What dropped benchmark performance?


o3 scores noticeably worse on benchmarks compared to its original announcement benchmarks


Any link / source / anything? You got quite an opportunity here, OpenAI employee claiming there's no difference and you got something that shows there is.


Yes, the original announcement for o3 and o4-mini:

https://openai.com/index/introducing-o3-and-o4-mini/

o3 scored 91.6 on AIME 2024. 83.3 on GPQA

o4-mini scored 93.4, 81.4 GPQA

Then, the new announcement

https://help.openai.com/en/articles/6825453-chatgpt-release-...

o3 scored 90 on AIME 2024, 81 GPQA

o4-mini wasn't measured

---

Codeforces is the same, but they have a footnote that they're using a different dataset due to saturation, but still have no grounding model to compare with


First post measures o3 at high reasoning effort. Second measures at medium reasoning effort. It’s the same model, then and now.


I've always found that these no-code workflow builders fail to hit the right abstraction - especially when new paradigms are added.


This is not no-code.

It allows you to write custom nodes with arbitrary code, but also connect them to existing integrations. Also, for each connection you can transform/select the data using JavaScript.


My leading theory is they're preparing for a increasing onslaught of spam "vibe-coded" shovelware


"I love sowing but I hate reaping"


Social Media suffered the same fate as all companies. A constant, relentless, unnatural pursuit of growth by stripping all humanity and focusing on numbers.

Social Media has turned into an unhealthy addiction


Untraceable and complete access to government databases. I can't begin to imagine the implications here.


We only hear about the cases where a someone is taking the risk of blowing the whistle, and actually manages to get the story out. Hopefully with enough substance for people to take the information seriously. How many cases that are likely to reach public knowledge is left as an exercise to the reader, as the saying goes.


>How many cases that are likely to reach public knowledge is left as an exercise to the reader, as the saying goes.

Is this some reminder to people that bad things occur that aren't found out.

Considering how everyone is aware of this is your comment some sort of clusterbomb whataboutism?


Direct access to private data relating to accusations against companies Musk owns.


This is the reality, unfortunately.

Lots of jobs that focus on communication and data organization are out the window, including recruiters.


They do need to train the agents first though, so that is the other purpose of agents, to gather the training data to replace the people.


Eventually agents from different providers will come into play. It's important to agree on a standard for accurate interoperability.

Ideally, the model providers would then build for the protocol, so the developers aren't writing spaghetti code for every small difference


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

Search: