jhonof

Wasn't Bell Labs this in the past? It existed alongside public research institutes.

Street parking in the core makes what would be a manageable amount of snow into a disaster like 2 weeks a year minimum because plows can't get down older streets.

I'm at 33 rejections 7 ghosts in my currents job search, is getting a rejection abnormal? I haven't really ever done a high volume search like I am currently doing so I have no context.

I genuinely use AI all day every day so it isn't just based on nothing. The PMs I work with have no interest in speccing the technical side of solutions out at all, and vibed specs aren't even close to being one shottable yet.

The argument is it drives down wages across the board so obviously the wages would be the same.

I disagree, I think being able to specify and validate a product is a massive skill that will be in high demand for a while yet.

Natural language is too lossy for most usecases and people are bad at being specific enough to get what they want. For example, I am not creative enough to think of movie ideas compared to good directors so I will probably just pay to see what they cook up rather than try to make it myself.

I am genuinely doing it for free in my free time on top of my work (in fact I pay for the models) so I am doing it for less than peanuts. I completely disagree w/r/t the model take, their is skill in steering work to get a product done that is what management has always been about, I do think in the long run this will be automated but I have seen 0 progress on it so far.

> There is no future in being a Proof Digester, no valuable human is motivated in doing that, nor is there any glory in it.

Coding has already gone this way and plenty of people do find motivation and glory in the final product vs. the building of the product (myself included). I absolutely see value in a person being able to humanize llm math output and see the field moving in that direction.

Silicon Valley was freaking out last democrat admin because anti trust laws were actually being enforced finally by Lina Khan, I actually completely disagree and I think this admin has been very obviously worse.

I actually heard the opposite, the folks at Anthropic felt pretty confident they were ahead after the Astra release because it wasn't as good as they were expecting.

Yes but at some point there is diminishing returns in model quality. Most people don't need a model that can solve Navier-Stokes to write emails for them.

Consent of the people can be different for different things? No way /s

Constantly ripping up trade/other deals that were signed in the last decade is unreliable and hard to plan around.

The login issue is extremely annoying, at least mandating an xcancel link would fix that.

It's bench-marking near sol

This is a succinct explanation for what I have been thinking as well, thank you for putting it into words.

I think the purpose of an event like this would be to optimize the process so that it isn't just occasionally asking an AI to keep going.

Qwen 3.8 27B is better than Sonnet 4.6

But they declared it...

We were already there with Opus 5 -> https://www.linkedin.com/pulse/nvidias-harness-hit-100-arc-a...

Yeah but that is a boat name, I don't think I've ever heard of a joke candidate in an election even do as close to as well as this against a high profile opponent.

Yeah I was thinking open router but I will look around at options, I genuinely think this model is good enough for like 90+% of my use cases, and the top frontier models are still not that good at architecture so I have to do that myself still so I won't be losing out.

Yeah this is the first model I have been able to run locally that actually feels useful, this is unreal I am considering cancelling my claude sub and going to just api (maybe GLM?) for really hard tasks.

28% of people voted for an alien trashcan helmet guy over Farage that's insanely embarrassing for him there is no positive way to spin this.

Isn't the US like very clearly anti free trade at this point?

Open source projects frequently have separate commercial licenses no? This isn't abnormal.

> For one, on the margin people are willing to pay a lot for slightly better models. I know personally the value the LLM adds to my workflow is considerably more than the $200/m I pay the frontier labs. I have no interest in optimizing that to get it slightly lower. There are a very vocal minority that optimizes this or companies whose LLM expense is marginal, but I think that's the minority (correct me if I'm wrong, curious what their customer base looks like)

Enterprise is how Claude/OAI make their money. Enterprise companies absolutely optimize around costs for infra, it's like an entire job category.

> Also the actual LLM is a tiny portion of the value added. Anyone that tried to build agentic solutions from LLM apis quickly realizes that a huge value is the Claude Code / Codex harness. There are open source implementations like OpenCode but they're not nearly as good.

Claude and OAI are not valued at $1T because of their harnesses

> Think about it another way. Consider how much money Microsoft spends on maintaining Excel. There are open source alternatives that have >90% of the functionality, they'll even work w/ Excel files and generate them. Google sheets is probably 99% and available to everyone and better in a lot of regards. But the immense value spreadsheet software produces workers above the $100 or whatever a year makes it so that there is a real moat and no one bothers exploring alternatives.

Enterprise excel is like $50 a month, the closed source labs charge orders of magnitude more than that per user per month for enterprise, and want to charge more if they can to maintain their valuation. I spent $20 in an afternoon yesterday using my work plan for one workflow for reference, I probably spend at least $50 a day and I don't even do that much coding anymore.

Why don't you just buy a chromebook or equiv and do that now then?

> Horizon can only be expanded at the cutting edges.

Ironically this was advice Ralph Waldo Emerson gave in 1840 so by your logic it's irrelevant just because it's old.