Jump to content

bargainman

Member
  • Posts

    1,136
  • Joined

  • Last visited

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

bargainman's Achievements

Community Regular

Community Regular (8/14)

  • Dedicated
  • Conversation Starter
  • First Post
  • Collaborator
  • Posting Machine Rare

Recent Badges

0

Reputation

  1. I'm kind of curious if anyone here is a developer using AI? Or in another field where AI, especially agents are making a huge push? Regarding "How much would you be willing to pay?", that's a reasonable question. But the ones to ask are, "How much does using this tool increase your productivity, and how much does that productivity cost/deliver in value?". There's nuance there, but if you're a software dev making 100K annually, maybe driving 250k in revenue and using agents increase your production by 12%, is 1k a month a bad investment? What if it 2x's your production? Now, does that 2x in production result in 2x revenue? probably not? Uber CEO seems to think that not. But it sure writes a lot of code. To me so far the biggest boon has been in 0 to 1 projects. I've created about 5(?) projects I would've never been able to even start. Did they increase revenue? Not all of them, plus I'm deeper down in the org, so there's no one-to-one. The actual production code is a bit harder to quantify. Teams still need to figure out how to get the massive amounts of code properly deployed, but that's a known bottleneck and will keep evolving. i.e. the models are really just the baseline right now. It's the agents and frameworks that are making the difference going forward...
  2. oh I don't know. A lot of those falls are kind of forced. Either way, not fun watching, and nba is an entertainment product... I'll take Nash any day over this guy.
  3. Yeah I'm not so sure. Actually it's less the random falling he does and more the hard elbows he gives the defenders as he drives into them that bug me. He's just not a fun player to watch it to root for. Not really a big fan of Luka or Wemby either. Wemby is just too tall it seems unfair I guess lol. I'm almost having to root for NY lol. I didn't like Brunson last year cause of his ball hogging, though this year maybe he's better. But it feels like if him and Kat win it's a nice story. For OKC this whole thing is really Kawhi Leonard's fault. Dude demanded Paul George and clippers gave OKC everything. Now they're set for years. On top of the new bargaining agreement which makes it really hard to build a top team any other way than the draft we'll be stuck with these 2 teams in the Western finals probably for a while. I wish joker had made it though. Nuggets didn't seem like a great franchise not building around him. He's super exciting to watch just for the great team basketball. I also wish LeBron had made it further, dude is incredible doing what he did at 40 plus OMG.
  4. Not sure who follows the nba. I like that there are Canadians representing.. I just wish the 2 most famous ones weren't SGA the flopper and Dillon the Villan. wish Nash were back
  5. You all remember Everlast? Just dropped a new song. Love this guy's stuff
  6. I'm curious, if you don't mind sharing, what is the nature of your business(es)? Were you looking to invest, ink deals, buy or sell product(?). Would love to hear details.
  7. Anyone else had their flight(s) cancelled by Westjet?
  8. You probably all knew this but I just learned Greg Abel Mr. CEO of BRK is Canadian!
  9. wow, so apparently the folks racking up 50 million views pushing fake news and trying to egg on the separatists are... Youtubers in Netherlands with no political conenctions, just trying to make a buck. Amazing. Of all the clicks to dollars paths to exploit...
  10. I will try to remember this next time. I'm not sure why it asks me so many times. It must think I'm robotic...
  11. woaw. I thought you had roots there but this must mean you live there!?? no one said anything about coffee crips, smarties, and Purdy's Hedgehogs yet!!! We visited Norway and survived. Although iirc we mostly got stuff from the grocery store and made it, can't remember a lot of out door meals other than the whale at Bergen's outdoor market. It's a beautiful place lots of sunshine in the summer! Reminds me a lot of the BC pictures you always see.
  12. I had a longer reply but then the site decided it wanted to check if I was human and killed my reply the number for the deficit was from the CBC video, AI just wrote it. The above is a great overview of Norway's history. Huge luck and advantages, and some good foresite. Also a small population helps. They have hydro and wind so pretty different from Canada's situation.
  13. Nice I was just about to post this. Basically sounds like a 'we'll backstop all of these investments that private money wouldn't go after on their own cause they aren't willing to take the investment risk, and we'll do it with borrowed money'. Norway is a totally different situation, that country is crazy rich, got really lucky, and they also made some really smart moves. Here's a quick summary from gemini The primary concerns center on the fact that this initiative differs fundamentally from a traditional sovereign wealth fund. Key downsides and caveats: Funding source (Deficits vs. Wealth): A traditional sovereign wealth fund (like Norway's) is fueled by surplus wealth, often from natural resources. In contrast, Canada is running significant annual deficits (reaching $78.3 billion recently) and high national debt. Critics argue this is essentially a "sovereign debt fund" built with borrowed money rather than surplus capital (2:49-3:53). Lack of Diversification: Traditional funds invest globally to maximize returns through compounding. The Canada Strong Fund plans to focus entirely on domestic projects, which limits potential for financial diversification and ignores the investment strategies used by successful institutional investors (4:41-5:36, 11:12-11:35). Risk to Taxpayers: The fund is designed to "de-risk" large-scale national projects to attract private investment. If these projects fail or do not generate the projected revenue, the government—and ultimately the Canadian taxpayer—may be left covering the financial losses (6:39-6:53, 8:33-8:52). Redundancy and Governance: Critics compare the fund to the Canada Infrastructure Bank, which faced calls for abolition due to concerns about whether its projects were selected based on political optics rather than financial viability. There is widespread skepticism that the new fund will repeat these issues without "ruthless insulation" from political influence (7:33-9:46, 12:45-13:25). Conflicting Mandates: The fund aims to achieve two potentially contradictory goals: acting as a national savings vehicle (wealth building) and driving industrial policy (nation building). Experts suggest that these objectives often do not go hand in hand, as projects that are "in the national interest" but ignored by the private sector may not necessarily be the ones that deliver strong financial returns (11:46-12:17).
  14. No comment on the guys. But that was in reference to training only. I'm not sure what the newer numbers are but training is a smaller part of the entire usage lifecycle. That's a fixed cost. After the model is trained it can be used for inference, and future trainings. Earlier numbers were 60-90 percent inference vs training. But with agentic AI I'm going to guess it will skew even further in inference's favor. Also note that 10x compute does not mean 10x the cost "The Vera Rubin NVL72 rack delivers approximately 10x higher inference throughput per watt" So the cost of 10x compute may be the same in a year as it is now for inference at least. The training costs per compute will also come down for the same models, but of course there will be more models. Agentic is really the difference. I know of people blowing through $2k tokens in a month and begging for more. They probably get paid 200k+ a year, so if the agentic token usage increases their productivity even in the low teens, it's a no brainer. Now that won't happen in all industries...
  15. Hmm I mean sure a lot of the data is garbage, but every model I've seen trained filters out all that crap data before the training begins. I'm not sure what he's talking about as if the models are actually training on that? Maybe they did when he worked at OpenAI? Kind of a strange comment. But perhaps it's because my work has primarily focused on SLM and on device so I don't see the greater slop? Anyway Karparthy is more knowledgeable than I am, and this is just a snippet so perhaps it's out of context (pun intended)
×
×
  • Create New...