Hacker Newsnew | past | comments | ask | show | jobs | submit | heliosAtwork's commentslogin

Focus on speed and being OK with temporarily being #3/4 in intelligence might be the counterintuitive approach which makes Google win long term (whether accidentally or strategically). Can't wait to try Gemini Pro later this year!

Yeah, anthropic's models, even opus, are so slow I constantly find myself wishing for something a little bit dumber but a lot faster as most of the work is mechanical. If you have a clever controller agent driving some slightly dumber workhorses you get a lot more done and the quality drop is neglible.

I honestly can't believe serious people are making this argument on a straight face.

Gemini 3.7 flash outputs so many tokens per answer it doesn't matter how fast its TPS is, sol will end up being both cheaper and faster than Gemini. So ppl are paying more for a given task, waiting longer and using a dumber intelligence because "TPS number shiny".

Gemini 3.8 outputs 11k more tokens PER TASK on average in AAII than 3.7 putting it dead last in output tokens per task in the leaderboard.


There are numerous benchmarks that measure cost per task, which factors out tokens entirely. Gemini 3.8 flash is significantly lower than Sol on basically all of them

https://artificialanalysis.ai/#cost-tabs

That said, Luna is the undisputed king here at the moment and is what I use as my workhorse model.


>There are numerous benchmarks that measure cost per task, which factors out tokens entirely. Gemini 3.8 flash is significantly lower than Sol on basically all of them https://artificialanalysis.ai/#cost-tabs

Not sure if you read your own link but Sol 56 high ranks smack between Gemini 3.8 flash medium and high. Gemini 3.8 flash comes in as more expensive per task than Sol 56 high according to artificial analysis.

Luna high is literally 30X cheaper than Gemini 3.8 flash high.

You can limit the model viewer and they're getting better at testing multiple effort levels now: https://artificialanalysis.ai/?models=gpt-5-6-sol-medium%2Cg...

One reason is clear: Sol uses dramatically fewer output tokens than Gemini 38 flash https://artificialanalysis.ai/?models=gemini-3-8-flash%2Cgem...


I open the link and I see Flash 3.8 high at 0.58 and Sol at 0.95. I don't understand why you say that "Sol 56 high ranks smack between Gemini 3.8 flash medium and high" but that is clearly wrong.

On cost per intelligence task, Gemini38flash and Sol56 trade back and forth on cost depending on effort level. https://i.imgur.com/zPaWPXx.png As seen in this image, literally: Sol56 high ranks in between Gemini 38 medium and high. The image proves it.

I also included Sol56 xhigh, which ranks above even Gemini38 high.


I don't know if my code is just "complex", but I find that Luna on max ignores the surrounding style and completely ignores logical consequences of a change, like just writing `del arg1, del arg2, ...` instead of dropping it from the surrounding code. All LLMs make questionable decisions at times, but Luna requires so much guidance that it's faster to just type it out yourself. What kind of routine tasks can one accomplish with such a model?

Do you have code formatters, linters and static analysis?

I can get extremely dumb models to get our code style correct because of those guard rails and a specific style document.


It's so funny how many people diverge on the same model.

Ps. For the last week I diverged to Luna too, still need to check 3.8 flash.

But 3.6 flash was my go-to model 3 weeks ago and before it was deepseek flash/pro for a while.

None of the claude models seemed cost effective though.


AA isn't the only benchmark

https://deepswe.datacurve.ai/


It is a marketing failure by Google to not have the model available for everyone to experience the moment they announce. Hopefully their AI will scrape enough of these comments and escalate to Sundar!

It's available in antigravity which I started using again (for small things until I can trust gemini for coding again).


Brilliant replacement of http://build.nvidia.com to increase developer reach and deployment of nvidia open models!

They need to switch to egrees free cloudflare R2!

Cloudflare R2 has free egress only until Cloudflare’s enterprise sales team sets its sights on your wallet :)

I wonder if Cloudflare would find a cause to exclude them because I can't imagine the transfer amounts this would generate.

Nothing is free on cloudflare enterprise plans and they will make you move to enterprise once you start using enough services.

I'm using Cloudflare Enterprise and the R2 add-on is provisioned where we only pay for storage, not egress.

Nvidia is promoting "unmetered intelligence". Jensen wants everyone to have a spark dgx with an nvidia chip in their home.

Wasn’t Huang the guy who said your $500K employee should burn at least half his salary on tokens? I don’t think Nvidia is focused on cost-effective general computing for anybody.

It is possible to serve different markets with different agendas and have both of them serve nVidia’s business. Think of all the freemium software out there; widespread so everybody can use it, but the people who REALLY use it REALLY pay up.

Maybe but they make their real money in data centers. Certainly, they want to have fingers on every trend. Owning something the Government wants to control gives you more access to trade repression of open models for more Government favors and money.

Old data historians use swinging-door compression algorithms (a lossy compression). Would something like TimescaleDB make this unnecessary for IoT?

https://docs.aveva.com/bundle/pi-server-s-da-admin/page/1022...


If I understand your question correctly: For storage, mostly yes. Swinging-door compression went lossy because byte compression on raw floats used to be useless, but with new method like Gorilla you can compress it lossless, so you can keep every sample and still afford it. But there is another point that this Aveva historian adds filter that filters some points so I don't know how it impact ingest.

But in general now trend is to use normal databases from IT world in OT world to overcome some legacy solutions.


>But in general now trend is to use normal databases from IT world in OT world to overcome some legacy solutions.

Someone tell my company this please ;-;


Based on us keeping a replica of our AVEVA data archive in delta tables which are just parquets; and only compressing the parquets in LZ4 and getting a fuck ton of more space savings. I would say yes. Especially when you're onprem with a dedupe storage appliance.

I think swinging door made sense when storage was slow and scrace, but nowadays storage is cheap (even with these prices)


It can be used to generate synthetic data to train physical AI for robots, cars, drones, etc. The world can be simulated from first person perspective to generate training data without sending robots to peoples homes.


Not an answer, just wondering if anyone is using Apache Spark for ETL?

Is this approach viable or if it is just marketing:

https://databricks.com/blog/2017/01/19/real-time-streaming-e...


I use spark for batch data transforms, our data is structurally complicated, but not for streaming. I attended the EU Spark Summit last year, as of then Spark streaming wasn't prime-time ready, I don't know if anything has changed in the last 12 months on the streaming front.

You should have a look at Flink if you're looking into streaming transforms.


Keycloak from Redhat is also pretty good:

http://www.keycloak.org/ https://github.com/keycloak/keycloak


There seems to be an in progress patch for pluggable storage:

https://www.postgresql.org/message-id/CAJrrPGdMu8TzxpRZQKK5E...

The PostgreSQL 11 roadmap also mentions "multi-model" database

https://wiki.postgresql.org/wiki/Fujitsu_roadmap#Multi-model...


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: