yipinwong

Say that to Luna's face. Ya all bringing up this not-so-cheap-nowadays chinese models and not that more intelligent than luna and bringing "cost" as the only factor.

One good alternative will steer me away from Google search like how Google took over Yahoo/Lycos/Ask Jeevs.

You spewing bs like the left-pad dev who removed it from NPM a decade ago crashing everything.

Yes, OSS can start with just a passion without a duty, but as the scale grows you have a duty to make sure it does not have adverse effect.

If you don't like the duty, just make it die by building it crappy like VIM does.

The topic isn't as interesting as how 131 authors communicated to get this out.

Just like companies are "legal entities" requiring registration, I got a feeling that the article is promoting the same for AI agents in the future.

GPT-6.0-Med LLC, Luna GmbH etc.

Regulations will come but this promotes anti-legal premise of "innocent until proven guilty".

i meant to say there is no such thing either, but your comment changed my mind. eye opening.

I can see where you are coming from. Vendor lock-in.

Basically spread many AI's everywhere, get people develop/user their ecosystem, and lock them in eventually.

But I find their models' intelligence lacking still.

I don't ever use their built-in gemini features (i have paid gmail) because they don't work well.

e.g. I ask gemini to format my Google docs per Google's material design spec with spacing, etc. It does a real bad job. Many times it does it line by line, and when I finaly get it do it for the whole doc, it does it sloppy, and extremely slow (takes 5 minutes for 10 page doc)

not as good as other chinese openweight models.

At best, good, but not great.

Any analogy works at certain abstraction level, and this works with a premise that it's a classifier that makes decisions.

Still good. In practice for Jev the devils in the details. As you all know by now, it's easy to write PoC and understand with AIs (or even manually, which is now a prestious practice).

That demo will get you 80% there

Getting to that 100% or even 99% to JEV level will be hard with all the edge cases, infra, API, communications, etc.

Still a good article.

Google is spreading too thin, as gemini isn't really that intelligent.

They are creating gemini SOTA (not really any more), flash versions, text-to-speech, video (omni), etc.

I can see they want to create an ecosystem, but I see no focus in any one area.

Price is a big selling point for a normie like me.

ty for the suggestion. I really never used "max/ultra" on Luna, and will give it a try.

mind your own business

Thing will die and go to grave in no time.

Benchmark doesn't really show the whole story.

I forgot which model degraded in quality as time went by, but let's try out Luna 6 for a few more days to confirm for upgradability.

I've been raving about Luna 5.6 as it's dirt cheap, and "intelligent enough". Double quoted.

Now GPT 6 Luna is even cheaper, and more intelligent, there is no going back... to SOL 5.6 for intelligent layer.

I am sorry, I meant to say I generated STAR out of my resume line, trying to generate STAR, and polish it thus $5.

---

I provided crapton of context for that one resume line. All the work I did, documentations for my justifications, etc.

I initially messed up and came out ot $5, rest of resume used around $4 per line (I used a fresh new session on purpose).

---

As a clarification, $2.2 average for OPUS 5.5 was the same process in a new session, same context, same prompts.

Also adding verification for that Fable 5.1 output in the same sesssion.

I spent about $5 per sentence in my resume using Fable 5.1 (High) to verify accuracy, inconsistency, and edit.

Opus 5.5 (med, as it's better than F5.1 high per graph in the article) used $2.2 and caught errors that Fable 5.1 missed.

Try Opus 5.5, cheaper, faster, and more intelligent for those prepping for interviews.

Not only that all DeepSeek is all garbage.

GLM or Kimi are better for my own personal projects. DS? uhm. it just keeps doing dumb crap

"legal" in media companies. It's a loophole in law where if you phrase something as a question, you are not claiming as truth.

Say the media claims, "Your mom could be the murder suspect?" You are not technically saying yo mama is a murder suspect though it is implied so.

Media companies won't be prosecuted for any "wrong info" for such titles by law because it's a "question".

Same here for the blog post. It's framing as a question, not taking a responsibility that (though author implies so) he is claming MCP is a bad idea wihtout owning it.

It's worse than weasel phrases.

Shady title with "?" not to own the statement.

Ending with a question is a shady media way of getting aroudn the legal issue of taking responsibility.

I hate blog titles like this as the author doesn't want to take responsiblity for the claim.

Given Jev does System 1 thinking, this would be equivalent to your ADHD heavy friend.

You are focusing on things that never happened or never will happen.

Phone is not the only communication you need for your kids or spouse.

For me it's not the phone that's stopping me from reading books but my computer, on which I play games, and browse, and do private stuff.

Where I live, electricity price is so ridiculous, I leave my PC off for 3 to 5 hours a day.

You get bored after 5 minutes. This is the best time to open your book and read.

----

Same idea applies for the phone. Just leave it off.

Bringing up iphone, I see how Jev pulled an Apple for making claims that their model is a breakthrough in research and 2 years of making like other iPhone features that's been around in other phones.

Baity claims worked didn't it for Jev? (most likely from AI forsure)

I might not have a good rep for Jev any more but at least I know what kind of model to use for decisions for graph engineering.

That's why I am a newbie for DB. I do not know how QO does that in the first place...

I haven't been to big talks. Only more like local engineer meetups.

A few times, I approached speakers for saying I liked it but turned down, as they had to talk to people from bigger companies, and few even expressed unpleasantness to me.

Those are rare cases, so I still give props to speakers whenever I can.

My lesson was that there are shitty speakers for sure, but don't let it stop you from continuing to appreciate speakers.

Reduce introduces state (accumulator), unlike map/filter which normally are used for immutability.

I use both, but do not like reduce at all. It's harder to read, yes. But I see the point of using them all.

Here is another "I love it" comment.

Genuinely giving the OP props