Jump to content
View in the app

A better way to browse. Learn more.

Surly Horns

A full-screen app on your home screen with push notifications, badges and more.

To install this app on iOS and iPadOS
  1. Tap the Share icon in Safari
  2. Scroll the menu and tap Add to Home Screen.
  3. Tap Add in the top-right corner.
To install this app on Android
  1. Tap the 3-dot menu (⋮) in the top-right corner of the browser.
  2. Tap Add to Home screen or Install app.
  3. Confirm by tapping Install.
Football ... Basketball ... Baseball ... Other Sports ... Futbol ... 🤫995🤫 ... Gambling ... Movies & TV ... Music ... Hobbies ... Lulz ... Food & Travel ... Daily Texan ... Business & Markets ... Cloak Room ... Help ... For Sale ... Board Discussion ... Advertise... Tailgate Donations

Captainant

Certifiably Surly
  • Joined

Everything posted by Captainant

  1. Captainant replied to Zavala's topic in Daily Texan
    I feel like we need some Aggy math to get there, they're experts at counting things that didn't happen
  2. I mean.....
  3. Agentic AI just means giving access to MCP servers and means to take actions in an environment - and none of that results in more revenue for openAI, but it does mean staggeringly larger context windows being passed back and forth in a transaction. From my seat riding shotgun, most "Agentic" workloads are just harnessing an LLM with some actions it is allowed to take. And even then, it's usually not giving consistent results that meet the business need. Hell, Salesforce is going through a rejection of AgentForce internally because of their inability to make it be more consistent in its behavior. MaarthandamSalesforce regrets firing 4000 experienced staff and repl...Salesforces has entered a phase of public reckoning after senior executives publicly admitted that the company overestimated AI’s readinessModels are too stochastic still, and frankly, most use cases and chatbots DO NOT NEED an LLM bolted onto it. Sure, a chatbot can theoretically handle greater variance in customer scenarios with an LLM backing it, but holy hell the business does not like when they get committed to something they can't deliver or when key steps are omitted because the agent just decided not to.
  4. IMO, this is the year of the bubble popping. I don't think any of these businesses are sustainable or viable, and it's going to negatively impact their MAU's and other metrics they're projecting ad revenue based off of. Brother, neither Google nor AWS nor Azure have enough capacity or ability to build out data centers to make the business model viable. And even if they did, the raw economics of that business model are upside down. In my experience as a technical expert in implementation and meeting a business use case, most people excited about AI as an investment and infinite money glitch don't actually use the technology in a deep manner. Yeah, they might get a shitty (and wrong) summarization from Google search, but that's not making use of genAI like its being pitched and sold as. Or even really scratching the surface of what it COULD do, if it weren't more expensive to run than any other application ever made. It's like all the biggest AI boosters (read: idiot executives) are just repeating what they've been told about AI and taking it at face value. And now that many companies are using it, they're finding that oh oops it's actually not that great, despite costing as much as a whole engineering team and consuming orders of magnitude more electricity and water than they would have used
  5. I mean, most of meta's advertising revenue is deeply fraudulent but that's a topic for another thread. (Link below is one example of one dimension of their fraudulent ad revenue) ReutersMeta is earning a fortune on a deluge of fraudulent ads,...Meta projected 10% of its 2024 revenue would come from ads for scams and banned goods.But let me put it into plain numbers for you: their userbase is on the order of 75-100M DAU's, and every single paying customer costs openAI more to serve than they receive in revenue. If they add ads, they aren't monetizing the paying customers that cost them money but they're still going to be pulling in $100-1000 per free user per month in advertising? That math doesn't math.
  6. We already have carte blanche for military installations and headcount on Greenland. I think we're about there, if they indeed start dumping treasuries
  7. It wasn't Hitler that hunted their neighbors and filled the camps and killed millions of Jews. It was the baker, the shoemaker, and the shopkeeper.
  8. It'll drag us all down with him. So that'll be fun, I guess
  9. Hey @ChickenSandwich what's the Ministry of Truth saying about this peaceful and lawful citizen being abducted and illegally held in violation of his constitutional right?
  10. The Jug is one of my favorite warbirds, thanks for sharing that deep dive!
  11. Even if/when Trump TACOs, the damage is done with our historic strategic partners and allies. Those bridges are burnt and will take generations (or post WWII style denazification) to rebuild the trust
  12. IMO, adding ads isn't going to generate the tens is billions in revenue. This isn't gonna help the business and it's mainly an optics play to keep useful idiots investing and to not pull their money out
  13. Captainant replied to Zavala's topic in Daily Texan
    Well I for one am glad the children are gone and the adults are in charge, right @realgreggym ?
  14. it was burning at 2 this afternoon, we should have heeded the ill omen
  15. We've already been there. Some models are better in different fields or tasks than others, but they're all just commodity at this point. The base technology of LLM's has been scaled lazily and greedily - in the technical sense - in that there haven't been many changes or advances to how the data is compiled, or models are trained, since the boom started a few years ago. Deepseek had the biggest step forward with applying a GAN architecture to train smaller portions of the LLM at a time rather than everything all at once, but even then the resulting model is still a commodity that is fungible with any other LLM. Even moreso if you're running your whole stack locally, since you really have no financial ties to the parent companies that build the LLMs. IMO, as someone who's been in the machine learning space since 2019, the majority of LLM service buying is being driven by idiot executives not wanting to look like they're falling behind. Not by any actual returns on the time or money. That said, there are actual good and interesting uses for the technology, but it's not what the trillions and trillions have been spent and committed on
  16. If only there were some warning, or trend of past business ventures that could inform us of such an outcome in advance!
  17. IMO the message was moreso that Mamdani is actually listening and getting results, and that cuts through the noise pretty well. The MAGAs still thought trump is doing the right thing, they just appreciated that they actually got something for their vote from Mamdani
  18. Captainant replied to Zavala's topic in Daily Texan
    Tucker really was the straw that broke the camel's back
  19. It's gonna be real interesting to see how this hits, if the markets are out of their k-hole and actually responding to events instead of just LGU
  20. Counterpoint: More Perfect Union did a really interesting deep dive on the strange newly budding "MAGA Mamdani" voter bloc Around 10% of Mamdani's voters are MAGA? Strange strange stuff
Football ... Basketball ... Baseball ... Other Sports ... Futbol ... 🤫995🤫 ... Gambling ... Movies & TV ... Music ... Hobbies ... Lulz ... Food & Travel ... Daily Texan ... Business & Markets ... Cloak Room ... Help ... For Sale ... Board Discussion ... Advertise... Tailgate Donations

Configure browser push notifications

Chrome (Android)
  1. Tap the lock icon next to the address bar.
  2. Tap Permissions → Notifications.
  3. Adjust your preference.
Chrome (Desktop)
  1. Click the padlock icon in the address bar.
  2. Select Site settings.
  3. Find Notifications and adjust your preference.