hamhamed

This is what I've been preaching to my team. With 5.5 and 4.7 the coding agents are good enough know to almost never take any tech debt. Any new feature or fixes should come with a cleanup or refactor, on the same PR.

It made the reds (coverts) way more pricier, so all is balanced (somewhat). see here, doubling and more in price: https://steamcommunity.com/market/listings/730/MAC-10%20%7C%...

At stay22 we've built a way to monetize travel videos automatically with multi-modal LLM, starting with travel and soon into retail. It's already live and testing with a few youtubers, via rev share. It automatically detects destinations, activities, hotels, implicitly throug the visuals or explcitily from the vlogger's voice

Few live examples: https://www.youtube.com/watch?v=A3tObRu8EuM https://www.youtube.com/watch?v=6v03mG1mMi0

That could be way at least for travel content if you work at OpenAI and want to collab

the em dashes are just giveaways that it's llm generated

Head of Engineering @ Stay22 | Montreal, Canada, Hybrid, Full-time | stay22.com

You'll be working with me (CTO) and leading a team of a dozen engineers to transform the way <scripts>, <a href> and <iframes> are done in the travel industry. We use a lot of AI (specifically traditional ML and LLMs) to predict, extract, and even do critical thinking for decision making when generating travel offers.

Apply here or directly to me: https://stay22.breezy.hr/p/b1215478578c-head-of-engineering

the irony is that the code is hosted on github, a microsoft owned corporation

I tried using langchain.js after a bit of hype and overwhelming dev support and incremental releases, but it became added complexity for no reason for our case.

OpenAI's NPM is much easier to use. A lot of what langchain promised out of the box didn't work like caching. Anyways it's still beta i believe

that tool should be a standalone service, something our team would pay dearly for. It goes through all slack public channel's history, creates embeddings and also goes into github, confluence, etc. And then people can just talk to the bot to get their questions answered if it can answer as good as chatgpt4

we work in the affiliate business and this has the potential to completely desotry the business model. Many of our partners rely on affiliate money to make ends meet, it is what powers most content creators.

Safari is planning to use ML to detect click_id type of query parameters and strip that from URLs. That's just poor execution and business destroying. PCM restrictions are horrible too.. we have to design the link so it stays within safari's specs:

> With an ad-click, an 8-bit ID can be transmitted (a number between 0 and 255, i.e. 256 possible values / campaigns) - per domain > For a conversion, a 4-bit ID is transmitted (a number between 00 and 15, i.e. 16 different types of conversion) - per domain

Not to mention Chrome and Firefox has other ideas, each different on how their PCM will be integrated. Other than the mega corps, noone is benefitting from this privacy enhancement. Just more work to adapt.

It is a statistical text predictor, and finetuning (RLHF) is the "magic" that gives it meaning and reasoning via weigths

Immediately came to HN, best status server

My personal recomendation for HN users to get into Ghibli is to start with The Wind Rises. It's more recent (2014), and loosely based on a true story/biography. No crazy monsters or spirits. It also emcompasses creativty, engineering and hard work as the main characeter design a fighter plane. Truly a masterpiece that resonated with me

Of course the other ones mentioned in this topic are good too.

We ourselves use BigML and AutoML from google, and we have engineers in our team who can do TF. It's just less of a pain in the ass, and the models are very reliable.

I'll give this a try, though I'm not sure what strategy it's using (I'm assuming an ensemble?)

Also you're asking the user if the objective should be classification or regression, I think that should be automatic, it's already buzzwords for your typical marketing person :)

GJ and gl!!

changing currency doesn't seem to work for me, stuck in EUR.

Anyways, this is great. Between lootbear, csmoney, and steam marketplace, I think i like your model best since it translates directly to cold hard cash. However, last I seen this model in the wild, it got banned.. re: OPSkins. How are you circumventing that?

We see around the same thing and our audience is totally NOT from HN or hacker like. Mixpanel and Cloudflare both report around 886k uniques (mixpanel being 5% less) for yesterday's Friday traffic... while GA reports 129k... that's a solid 80%+ too.

I'm not sure why everyone is tripping here. He's about to save some fucking lives and you suddenly care about the enviroment for sending a mere 100 ETH?

Even if we are to look at the big picture, don't tell me you believe PG will be the sole reason everyone is going to pick up NFT donations now? It was going to happen regardless.

Totally agree - you can either spot those yourself by putting yourself as a user or listening to your customers. The other challenge is to get those tiny wins into the sprint when your PM can be clueless about the ROI on those..

While we're on the subject, if someone from Mac OS can please seperate my trackpad and mouse settings so the scroll natural/reverse is kept between for each device, that'd be appreciated

I've been spending almost every weekend for the past few months trying to break a crypto crash game (called bustabit) using ML. While I'm also dealing with probabilities, my challenge has to do with the limited data I'm dealt with even though their API is public.

In horse racing, you have access to so much shit that you can feed into the AI, like rider statistics, horse performance, horse attributes, etc. In my case, I only have the game/round number and the bust result (red or green). From thereon, I have to create many other variables (aka featuring engineering) like round win/lose streaks, moving averages, MACD, etc. I can also feed in what and if other players are currently betting for the round or not.

Currently, the house always wins. And their game is rigged so they have that 1% advantage over you. However, so far with my AI, I've came with enough confidence score to win as long as my bot plays 0.001% of the games.. which isn't declared victory since those blackswan games happen once every 3 weeks.

There's also another way to beat it without using any ML, just plain old fibbonaci martingale - requires around 10k starting money for error margin but it ends up always ahead of the house.. That's just a math way to beat it and not very fun :)

It's probably crazy of me to even attempt to predict randomness, but there's something intriguing when you mix big data and ML with forced randomness. Again, they have to keep forcing the odds to be 49% vs 51%..

Thanks for the free service, email getting leaked is the least of my concerns. Can't get mad, thank you Dan. HN should have this natively

I don't know about abundance. Maybe Aaron has confirmation bias being around the industry so closely. Or maybe other founders here can tell me otherwise. From my side, I can't vouch for an over abundance of capital.

It's been a year we're trying to close our round, missing 350k from the $1M round and yet still growing 30% MoM. Yes covid fucked our business in a sense, and we got hit since March - but we remain ever so bullish esp given our sustaining growth.

Youtube link if you want to speed it up https://www.youtube.com/watch?v=Unzc731iCUY

Stay22 | Montreal, Canada | Frontend Engineer (mapping/embed) | Fulltime | ONSITE Stay22 (founded in 2016) is one the fastest growing startups in Montreal that helps organizers improve the overall experience of their participants through an embeddable map that makes it easy to search for Hotels & Airbnbs near an event.

This is a senior/lead position: experience with either mapbox, or google maps, or leaflet, etc. is a big plus.

- Developing front-end web applications that perform well on all devices, always maintaining a decent frame rate. - Familiarity with performant vs non-performant CSS rules and properties - Knowledge of browser internals, how it draws and paints, how HTTP works, etc. - Lives and breathes vanilla JS and the DOM.

Reach out to me at [email protected]. You will be taking part of my job :)

well, there's constantly empty chairs on the face cam whenever it's on the OpenAI hero, so something must be fishy ;)

Pain isn't the toughest of opponents at TI. Will they be playing vs a team that is currently still alive in the bracket? So far it seems Pain is kind of controlling the game at least much better vs the caster team (99th percentile)

We have raised, around 700k, with a core team of 15. And every month has been our best month for the past 8 months. As we gear up for our A round (or US seed round) we'll try our best to maximize our chances to get in front as much investors as we can and YC will def be a great bonus for this part

I'll def reach out once our application is done, we never considered getting it reviewed before, so that'll be a first..thx evan!

And yeah, as you may have guessed wedding platforms like Joy, Knot, etc. are def in our target channel

I've been applying to YC since 2013 (so this is going to be around our 10th time), although with different companies and cofounders. I've never gotten accepted, though we've made it to the interviews twice but never made it to the finish line to actually get in. Here's what I learned that could mean something:

1: Every time we got an interview we asked for recommendations. Last time we got a straight up email rejection when we decided it was probably a bad idea to ask again our network for a recommendation for the third time. If anything our startup was doing better than the last time we applied so I'm not sure how their process works.

2: We recycled our 1 min video and we got an interview. So if you're too lazy to make a new one, I guess it's fine

3: We'll apply for the 10th time sometime this weekend, and probably for the 11th.. fuck it.

What would be the requirements to get into this program? I.e team size, previous amount raised, etc?

Stay22 | Sales and Partnerships | Montreal, QC | Full-Time (Onsite or REMOTE) | https://www.stay22.com/

Team of 7 based in Montreal transforming the way event goers book their hotels and short-term rentals. We've partnered with many large events and ticketing platforms. Stock + Commissions + Salary

Example event: https://en.esl-one.com/dota2/hamburg-2017/visitors/

Email me at [email protected]