HNHacker News
TopNewBestAskShowJobs

griomnib

739 karma · joined November 11, 2024

submissionscomments
griomnib··on Google's Results Are Infested, Open AI Is Using Their Playbook from the 2000s
It’s just not possible. You can do a lot with nondeterministic systems, they have value - but oranges and apples. They need to coexist.
griomnib··on Carlsen disqualified from World Rapid and Blitz championship for wearing jeans
Totally agree.
griomnib··on Google's Results Are Infested, Open AI Is Using Their Playbook from the 2000s
But they can never be. RAG gets you somewhere, but it’s still a pile of RNGs under a trenchcoat.
griomnib··on Google's Results Are Infested, Open AI Is Using Their Playbook from the 2000s
I think when it’s good it’s pretty good.

But knowing when it is good is still hard, as I can’t trust it more than an LLM. But with an LLM I have a simple chat window, not a bag of rabid SVPs fighting to be on the SERP page.

griomnib··on Carlsen quits World Rapid and Blitz championship after dress code disagreement
I find people who use it in a derogatory don’t do much actual thinking to start with.

The literal origin of its use in social justice realms was to put an emphasis and value on critical thinking.

By pondering the nature of society one can be “awoken” to the deeper sociological forces which shape individual lives.

I’ll freely admit the term has been perverted by a range of actors with motivations I disagree with.

But the original meaning and intent is valuable: all of us live in a world of hidden boundaries and power structures, but seeing that requires thought.

Many people don’t want to think for themselves, and they hate those that do.

griomnib··on Google's Results Are Infested, Open AI Is Using Their Playbook from the 2000s
Google was originally fairly egalitarian, OpenAI never was, and never will be. For better or worse.
griomnib··on Google's Results Are Infested, Open AI Is Using Their Playbook from the 2000s
LLM are non deterministic by nature.
griomnib··on Google's Results Are Infested, Open AI Is Using Their Playbook from the 2000s
Funny thing is you can train a small BERT model to detect queries that are in categories that aren’t ready for AI “answers” with like .00000001% of the energy of an LLM.
griomnib··on Google's Results Are Infested, Open AI Is Using Their Playbook from the 2000s
They have something Google never had: a paid tier.

There are plenty of revenue models aside from ads.

griomnib··on Google's Results Are Infested, Open AI Is Using Their Playbook from the 2000s
The biggest reason SEO is profitable is because low quality sites run display ads. That is the lifeblood, and intrinsic motivation, for these sites to even exist.

Google operates the largest display ads network. They literally *pay* websites for SEO spam, and take a very healthy cut off the top.

I wish people would stop acting like Google has been in a noble battle against the spam sites, when those sites generate Google billions of dollars a year in revenue.

The obvious question is, why would they ruin search for display? The answer is greed combined with hubris. They were able to double dip for years, but they killed the golden goose.

Everybody with a brain knew this would happen when they bought Doubleclick, and it took longer than expected, but here we are.

griomnib··on Veo 2: Our video generation model
No they’ve been acquiring layers upon layers of middle management for a decade.

That’s the core issue, and they’ve also pissed off a non-zero percentage of top talent by ditching what still existed of Google culture and going full “Corporate Megacorp” a few years ago.

Google is having to pay a ton to retain the talent they have left and it’s often not enough.

griomnib··on Veo 2: Our video generation model
SD will also generate what I tell it, unlike the corporate models that have all kinds of “safeguards”.
griomnib··on Veo 2: Our video generation model
Or, using Occams Razor; Sundar is a shit CEO and is playing catchup with a company largely fueled by innovations created at Google but never brought to market because it would eat into ads revenue.

That, or they have a secret super human intelligence under wraps at the pentagon.

griomnib··on Veo 2: Our video generation model
They already did it, and I’m guessing they were using some of the various YouTube down loaders Google has been going after.
griomnib··on Veo 2: Our video generation model
“Won” what exactly? I have no issues running stable diffusion locally.

Since Llama3.3 came out it is my first stop for coding questions, and I’m only using closed models when llama3.3 has trouble.

I think it’s fairly clear that between open weights and LLMs plateauing, the game will be who can build what on top of largely equivalent base models.

griomnib··on OpenAI whistleblower found dead in San Francisco apartment
Sure seems like this is happening more frequently, eg with the Boeing guy. So it’s reasonable to ask why.

If you look at Aaron Schwartz for example you see they don’t have to assassinate you, they just have so many lawyers, making so many threats, with so much money/power behind them, people feel scared and powerless.

I don’t think OpenAI called in a hit job, but I think they spent millions of dollars to drive him into financial and emotional desperation - which in our system, is legal.

griomnib··on macOS 15.2 breaks the ability to copy the OS to another drive
It appears “solid”, but the only test of a backup system is disaster recovery and I’ve had to attempt to get files from Time Machine drives before and it is hell.

iCloud Drive is also a bit iffy, it does a lot of weird shit with file meta data that can screw up file processing.

griomnib··on macOS 15.2 breaks the ability to copy the OS to another drive
Apple’s built-in backup solution is called Time Machine. It was dogshit when it launched a decade ago, never got more reliable despite many OS releases, and is still largely dogshit.

The author of the blog makes an application that is very good at making backups on macOS and now their app is broken due to skullduggery by Apple.

So the issue is threefold: a) Apple further turning MascOS into iPhone, b) this harms competition, and c) users ultimately suffer because Apple has no incentive to get better.

griomnib··on Carta is making it too difficult to cancel subscriptions, some founders say
Realistically the FTC doesn’t have the staffing or expertise to act on much at all.

They can offer “rules” of questionable value post-Chevron, but even in areas like COPPA where they have clear statutory authority, and widespread lawbreaking, they do nothing.

However, like all things in the United States, justice is for sale, you just need to sue.

griomnib··on BlenderGPT
Depends if you count violating data protection law is a crime or not.
griomnib··on Making “social” social again: Announcing Mozi
Queers always lead the way, without queers bluesky never would have taken off
griomnib··on Making “social” social again: Announcing Mozi
“I realized I had spent too much time on startups and not enough on building relationships. So I decided the only way to fix it was another start up.”

Huh?

griomnib··on Gukesh becomes the youngest chess world champion in history
Hey now, that microservice happens to violate data protection law in 40 jurisdictions per second, that’s basically a criminal mastermind.
griomnib··on BlenderGPT
Oh, if they had YC backing I’m sure half the people here would be sending in job applications.
griomnib··on BlenderGPT
They aren’t software, Blender foundation likely has exclusive use in that domain. I’m guessing if it was litigated it would come up, but that they’d prevail.
griomnib··on Gemini 2.0: our new AI model for the agentic era
Ok I tried literally last week on Pixel 7a and it didn’t work. What model do you have? Maybe it requires a phone that can do on-device models?
griomnib··on Gemini 2.0: our new AI model for the agentic era
And where is the ship headed if they are no longer supporting the open web?

Publishers are being squeezed and going under, or replacing humans with hallucinated genai slop.

It’s like we’re taking the private equity model of extracting value and killing something off to the entire web.

I’m not sure where this is headed, but I don’t think Sundar has any strategy here other than playing catch up.

Demis’ goal is pretty transparently positioning himself to take over.

griomnib··on Gemini 2.0: our new AI model for the agentic era
Latency is a huge factor in performance, and local models often have a huge edge. Especially on mobile devices that could be offline entirely.
griomnib··on Gemini 2.0: our new AI model for the agentic era
And yet these new models still haven’t reached feature parity with Google Assistant, which can turn my flashlight on, but with all the power of burning down a rainforest, Gemini still cannot interact with my actual phone.
griomnib··on GM exits robotaxi market, will bring Cruise operations in house
The competition is human drivers.
← PreviousPage 4 of 8Next →