larodi

So it goes like this: I first you take, then money you make and eventually you repay. This is a very smart loan from society indeed. And seems to be the new normal…

Even though Anthropic is my daily driver I’m done respecting any sort of copyright. I’m okay paying for subscription for a service delivery but never ever again will I believe in copyright or any other utterly non-enforceable similar concept.

And no similar sized big model is even public, which is super important to note.

The name for it is ablation - precise removal of parts of the model. Not abliteration as it is not obliteration.

Even as I write this the ‘abliterated’ word is denoted a typo. Does it not at your end?

Tis surely Twain, but come on, he came ages after Homer and came with civilisation that was not there in America when Homer lived.

The whole idea for looking for America’s …Homer is just super bizarre. Modern (last 150-200 years) of America is not what it was thousands of years ago - it stands on the burial grounds of millions of indigenous people obliterated altogether with their cultural inheritance.

It is questionable even today if American authors bear more the heritage of European literature than anything else.

Homer was 2k years ago. Conquistadors made sure nothing would be left of anything that was 2k years ago. English and French, besides being smarter in the north, dis likewise.

So your American Homer’s legacy perhaps lies buried down in some unmarked indigenous grave.

Ok, so let me wrap this for everyone:

While everyone here in this forum kept arguing (and fighting and yelling at each other) whether tis moral/right/secure/cheap for ppl to rewrite and ship a major software package with LLMs/agents, one of the main drivers behind AI actually did it, without consulting your opinion, and most same everyone actually slurped this decision without paying a a notice.

Boring is good, but this boring is super massive major thing that happened, and precisely because security is still intact.

So, tis for leets, not for folks…

It is not the model now showing intelligence. Tis the dialogue. The medium is the LLM and the harness is then a next level construct.

Surely not that good at vision. TBH none of these 14-27b models come close to even the cheapest Gemma 2.5 flash.

If these buddies are similarly bad on text, then they definitely don’t get anywhere close to big boys, no matter what the synthetic stats claim upon release.

Because you not even trying

supercarblondie.com

Published article bout supposed breakthrough battery recovery with zero refs to actual sources as the other commenter noted.

So likely the article is LLM reprint of something else or aggregation. Thus blondie slop.

But even so it tops like crazy because … well because, in all fair evidence, HN folk seems to care zero about sources, or less so about authorship when is/gets excited about random shit, but in the same time cares badly whether sacred code in Linux kernel was AI written.

Now downvote more.

Just got downvoted few weeks ago for suggesting someone would soon rewrite Linux into Rust with LLMs and now he did, and this is topping HN.

Are you guys the same guys who read HM every day or some other completely new folk?

Claude vision seems like a massively overpriced VLM choice given Gemini has great VLM tower since 2.4 and my gut feeling is this is a task that local Qwen’s would handle easily.

Disclaimer: my daily workflows include millions of Claude and Gemini tokens spent on dev and video vision intelligence.

I can sense Claude behind all interactive graphics here, yet it apparently does great job to further make the point and interestingly TimesOfIndia is a major media outlet.

If you're a job seeker - create a prototype as close to a problem that the target company may be solving - call them, and show them. Works a charm.

Vision changes are part of any intoxication, most avid alcohol users have them all the time. You have vision changes due to high blood pressure, so vision is perhaps not a KPI to put on the table as a major one.

I’ve had it three times already in the last 2 years and it is something real. Dizziness and depressive swings is the first that comes, then comes memory loss and finally trouble with speech.

At this point you should stop risking to burn your central cognitive capacity. Be advised.

Note: myself a passionate Claude code user with multi agent parallel approach to dev. Plus 30 yrs of various oldskool dev experience. My blood and sugar and all tests are all within norm, and I bike and swim regularly. I’m not a major drinker and try to avoid alcohol in general. I count 25 trees outside my window and I’m not on any amphétamine pill such as Adderall.

And then Dutch-Americans leave for Bulgaria and Romania .) occasionally Croatia

quote from the page:

"This page ranks 42 party size Bluetooth speakers. If you’re looking at one of these, you need very deep bass or extreme loudness and can accept a large, heavy speaker. Most are overkill for typical use, the Boombox size is better for most people. Most are at least IPX4 unless noted. All need AC to charge. The best here have bass extension similar to a 10” woofer tower speaker."

my note: page has many links to youtube tests, and resources, does not seem to be advertising-related, though is biased towards JBL.

and it streams back in time to 1998...

Indeed I suppose many are on the same opinion here… save for some super stubborn uneducated jerks who keep downvoting all my ‘Alcochol-is-likely-worse-than-LSD’ comments without even getting into reasonable argument.

This is absolute nonsense and is not backed by any sane scientific evidence. Using alcohol even once, in the quantities readily made available everywhere where alcohol is available, can finish you off in the most diverse and bizarre ways.

What’s even more ridiculous is the fact even quarter pill of the Xanax kind mixed a very limited amount of alcohol gets you potentially higher than most psychedelics when properly dosed. And both opioids are super well available…

There’s more evidence it is than it’s not. And, yes, even people who have been super stubbornly against psychedelics only very seldom come to regret taking some. Which you cannot say for other recreational or sedating stuff at all.

According to the GPTs, the LSD ranks the bottom in the chart of ‘damage done to your organism’ while alcohol tops it every time.

So are you trying to imply that I've somehow accidentally stumbled upon more than 100k lines of new high-perf working code, done in less than 6 months, which is like not 20%, but 200% my actual output, and this code, already generating revenue for me and my employer, is something of the ordinary, and actually I can do 20% better typing it manually?

ROFL sorry

I guess there aren’t so many people here who can proudly say their mom used to program MUMPS while they were still a fetus, but well that’s me.

Its a different planet entirely. At least since Jacquard's loom ran the first program. Perhaps even earlier - when the printing machine did the first print.

Well… tis difficult if one does not understand how grammars work, and therefore parsers. But we’ve seen people use stuff like ContextFreé’s Design Grammar, without even being IT guys, and still figure themselves around.

The whole notion grammars are hard is just wrong. They are not only powerful, but super simple in fact. As is the basic regexp if one cares to spend a focused afternoon to understand it. Probably even less time if working with a decent teacher.

Font claims to support Bulgarian Cyrillic but in fact supports the Russian one, which is different and very often overlooked as the same by font designers.

I would prefer GroundingDINo which is a sort of SAM and Dino combo which does open vocabulary.

One thing I don’t get I why the article is credited to ‘Contributing Author’.

Meanwhile their very own Peter Skalski already does super job with host write ups and examples of all YOLO sorts and is well respected.

Read through his papers, and these are substantial accusations. Perhaps Hinton et al. should investigate, and either a) correct themselves by properly citing these Munich researchers; or b) proove they did not base their work (unlikely) on these papers;

And then, as a whole, this weighs in favor of European scholars and also should properly inform the funding of similar research in the EU.

Writing the last in the light of a month-and-a-half wait (to date) for EuroHPC to process their own form where we submitted a funding request by no less than University + Private Company already established in the area + 4 alumni, two PHDs and one postdoc. Zero response since.

Some if not most coverage will be off, indeed, but then the important stuff can get you lots of benefits. This makes sense even today for selectively patching the kernel. I’m sure many people been odd by the complexity of it while now it is doable albeit with agents…