HNHacker News
TopNewBestAskShowJobs

CSMastermind

9,141 karma · joined July 9, 2014

christopher.seltzer@gmail.com 7246304589
submissionscomments
CSMastermind··on Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
All the benchmarks I see put it around the capabilities of Opus 4.8 Medium or Sonnet 5 High.

As far as I can tell it's slightly better than GLM 5.2.

CSMastermind··on Microsoft deleted my account and OneDrive
Back in college, I did a hackathon where I striped data across different clouds like OneDrive and Dropbox so that even if one of the providers went down you would still be able to recover your data. Had the additional benefit of no single provider being able to read your files because none of them had the full data set.

Ended up not pursuing it but every time I see a story like this I think I should have.

CSMastermind··on Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k
Everything about LLMs is inefficient. They have their benefits but watching them reason over things that are painfully obvious, that they've literally investigated before (before a memory compaction), never take a step back aand be like 'this is going too slow let me look for a better way', etc. is painful.
CSMastermind··on GPT-5.6 Sol Ultra produces proof of the Cycle Double Cover Conjecture [pdf]
There are different classes of mathematical problems.

The one in the post definitely shows the advantages that LLMs have compared to humans for some problems but it's in an entirely different class than the Riemann Hypothesis.

Riemann is one of the most studied math problems of all times and all of humanity has basically collectively failed to make progress. The idea that there's some technique that just hasn't been tried yet (like in the post) is very very unlikely.

The general consensus is that we'll need an entirely new branch of mathematics to solve Riemann - our current tools aren't just inadequate; they're of the wrong class entirely.

I suspect inventing new branches of math will remain beyond LLMs for the remainder of my life.

CSMastermind··on New York City to ban deceptive subscription practices
Sure but say some company based in Lousiana has a website that violates NYC law.

What exactly will NYC do about it? Is there some mechanism to for them to block the website inside of NYC? The company would presumably have no property they could seize or employees they could imprison.

If it were a state passing the law then they could sue for enforcement in federal court but I don't think a city could?

CSMastermind··on New York City to ban deceptive subscription practices
Is this something that cities can really enforce? Like I get that NYC is a bit of an exception but let's say a 5 person town in Wyoming decides that they want to make this practice illegal and they all vote to do so. It's not clear to me that would mean anything at all.
CSMastermind··on GPT-5.6 Sol Ultra produces proof of the Cycle Double Cover Conjecture [pdf]
I mean if there's something I'd bet against being solved by LLMs in my lifetime it's that one. We truly do not have line of sight into what a proof would even look like.
CSMastermind··on Google Earth Pro desktop downloads will end in 2027
I despise Google's insistence that everything be a web page.
CSMastermind··on ChatGPT Work
> expanding it beyond coding work

Is there anything it can do that codex can't?

CSMastermind··on ChatGPT Work
How is this different than Codex?
CSMastermind··on Grok 4.5
I've seen Opus do this pretty frequently, actually. It's one of the reasons I don't trust it as a model.
CSMastermind··on Separating signal from noise in coding evaluations
DeepSWE is the one I generally trust: https://deepswe.datacurve.ai/
CSMastermind··on The great blogging collapse: What happened to 100 successful blogs?
Substack is doing just fine. Blogging didn't collapse, a bunch of spammy get rich quick types were a flash in the pan as expected.
CSMastermind··on EFF letter to FTC on X consent order [pdf]
Reddit did the same. Tumblr died when it banned porn. There seems to be a very perverse incentive for social media platforms to be as permissive as possible.

Personally, I'd be in favor of banning all sexual content on X, but it really feels like a legislative solution applying to all social media platforms might be the best solution.

And yes I realize the slippery slope that could put us on.

CSMastermind··on Senior SWE-Bench: open-source benchmark that assesses agents as senior engineers
Man I don't know if I'm living in a crazy bubble or something but GPT 5.5 is lightyears better than Opus 4.8 for me to the point where I'm honestly wondering how you're evaluating them or what kind of work you're doing.

There's specific tasks that Opus does better on like Frontend Dev and Design but for anything else 5.5 just laps it.

CSMastermind··on Claude Sonnet 5 – benchmark results
Using Fable, pretty much every request hit some gate they had for no discernible reason. These provider-level rejections should be incorporated into benchmarks as 0s on the tasks since that's the experience you'll actually get using the model.
CSMastermind··on How H-E-B became Texas' most beloved brand (2024)
They used to have HEB, Wegmans, etc quality stores in Seattle before some merger? What were they?

Why did this M&A seemingly not happen in Texas, New York, Florida, Pennsylvania, and Ohio?

CSMastermind··on How H-E-B became Texas' most beloved brand (2024)
I have a weird affection for good georcery chains. Wegmans, HEB, Giant Eagle, to a limited extent Publix. It's so odd to me that there are areas of the country with nothing like them. It really seems like you just need one good family who decides that they're going to care about how they're set up and make an effort to build a company with the right values.
CSMastermind··on Previewing GPT‑5.6 Sol: a next-generation model
5.5 is smart enough for 99% of my tasks. I need that level of intelligence at ever decreasing prices.
CSMastermind··on Claude Tag
Lots of people use design
CSMastermind··on Codex logging bug may write TBs to local SSDs
Lol this was my theory as well.
CSMastermind··on SpaceX to buy Cursor for $60B
I use grok every day as my go to search engine replacement.
CSMastermind··on SpaceX to buy Cursor for $60B
> I was switching between Opus and GPT 5.x and was spending anywhere between $500-1000/month.

I switched to Codex because Cursor was costing me $10k / month. If the price goes up by 4x I'm not sure the value will be there for a lot of people.

CSMastermind··on Microsoft turns to AWS as GitHub faces AI capacity crunch
Doesn't most of Anthropics revenue come from AWS still?
CSMastermind··on There is a shadow hanging over this Fable thing
> OpenAI did the same “too dangerous to release” song and dance for the awesome, world ending AI that was GPT-2.

Wasn't that when Dario, et al were at the company. One way to view this is that OpenAI expelled the cultists and they went on to form their own organization that continued using the same tactics.

Certainly some of the Anthropic press around Fable seems to me to be just marketing but I also think there's a core of people there who really believe it. I also think like all good advertising/lies there's some truth to the claims even if they're exaggerating.

CSMastermind··on Statement on US government directive to suspend access to Fable 5 and Mythos 5
> I find it worrysome how often people value revenge over good.

I personally see it as a net good if companies fearmongering for marketing purposes then have to face consquences from people taking their marketing at face value.

Hopefully it teaches them and others not to do it anymore.

CSMastermind··on Statement on US government directive to suspend access to Fable 5 and Mythos 5
Their entire marketing strategy has been unwarranted fearmongering. This is completely unsurprising.
CSMastermind··on Waymo Premier
Their coverage area in Orlando is so limited I functionally can't use the service. Would be nice if they could expand it to at least include the airport and tourist areas.
CSMastermind··on Anthropic apologizes for invisible Claude Fable guardrails
They should apologize for their visible gaurdrails, I don't think I've had a conversation that hasn't downgraded to Opus for completely inexplicable reasons.
CSMastermind··on Who's the smartest corvid?
I'm not an expert in the area but have read a bunch on this topic to try and understand it better. Bird brains and human brains are structured very differently. Birds are much more like GPUs with independent distributed processing happening in parallel. Mammals have these big bidirectional layers where signals are constantly propagating up and down in a big connected computation.
← PreviousPage 3 of 34Next →