Marko Cadez

[Saganism @Saganismm]

A physics professor summarizes the essence of quantum mechanics in a minute.

[Marko Njegomir @njmarko]

@cb_doge The actual cover should have credited the people whose research led to this AI boom @SchmidhuberAI

[Marko Cadez @cadezmarko]

@BoraKonj Evo mojih predloga: Vi imate uticaja da im kažete bar nek prave kako treba - a ne ona jakna pako i frajer krš

[Marko Cadez @cadezmarko]

@samodejo01 Budite malo kreativniji aman🤣 makar mi skockajte potencijalnog dečka i obicite mi nesto pristojno: ovako nesto….

[Marko Cadez @cadezmarko]

Quote of the year or decade or millennium….. let's see

[Marko Cadez @cadezmarko]

Check out the latest article in my newsletter: Nicolas Mottis: "If we innovate only for profit, we risk building empty shells. If we innovate only for values, we risk lacking impact." https://t.co/gClG6mZbNi via @LinkedIn

[Marko Cadez @cadezmarko]

It is Sunday. Time for religious practices so….
DAVID GUETTA, HYPATON & BONNIE TYLER - TOGETHER https://t.co/VsI6nSZCuo via @YouTube

[Gert Jan Koopman @GertJanEU]

Great exchange with business leaders at @wb6cif's 🇪🇺EU-Western Balkans Business Platform on the region's investment potential.

Our 🇪🇺#GrowthPlan brings new momentum for early EU single market integration.

We'll step up our private sector engagement to unlock more opportunities!

[Privredna komora SRB @Privrednakomora]

@cadezmarko Delays at the European Union borders cost companies from Serbia, the region and the EU €130 million annually. @RTS_Vesti

[Marko Cadez @cadezmarko]

@aleksabgd1 @TPicula @Europarl_EN @TheProgressives Pitali, uputio nas na Nemacku, Austriju, Francusku, Italiju….i mnoge druge zemlje EU koje imaju zakonsko (obavezno) članstvo u svojim nacinalnim Privrednim Komorama. Koristite naše usluge ukoliko do sada niste https://t.co/jwXdp6txqp

[Tonino Picula @TPicula]

Sastanak s Privrednom komorom Srbije. Razgovarali smo o ekonomskim kretanjima u zemlji, ekon. odnosima s EU i drugim zemljama te kako iskoristiti mogućnosti koje pruža Plan rasta za Zapadni Balkan, na kojem sam radio kao suizvjestitelj @Europarl_EN @TheProgressives

[Marko Cadez @cadezmarko]

Love it ❤️

[Dejan Vuckovic @DejanVuckovic12]

Marko Čadež ga kaparisao za otvaranje EXPO 2027.

[Magna Ding @MagnaDing]

OpenAI drops ChatGPT's mind-blowing conversational chat update.

It reads audio, deciphers video, and reads emotions in your voice.

Here are 5 examples you should check:

[Marko Cadez @cadezmarko]

@USAmbKosovo @USEmbPristina @odaekonomike @Privrednakomora Your Excellency thanks for this opportunity to address the main challenges for our business communities.

[Marko Cadez @cadezmarko]

Brat Djidji!

[Čedomir Jovanović @Ceda_Jovanovic]

gospodin starog kova, smeker iz ulice Vase Mickina, Stokholma, Azurne obale…pamtim da smo se upoznali tako sto sam sa bine na Trgu republike u masi video coveka koji me dok sam govorio o ratovima, podeljenim narodima, izgubljenim zivotima uplakan slusa dok mu se suze slivaju niz lice. Kasnije smo se upoznali i tesko da ce nas za zivota razdvojiti. Kada me on zatreba u Akri ja “resim” za pola sata, kada se ja detoniram on sve primi na sebe…dobro, ponekad pozovem Halida Beslica da ga opomene, jer Cola i Brega su mu raja

[Google Labs @GoogleLabs]

Lab Sessions are collaborative AI experiments with visionaries across disciplines. In "Journalist x LLMs," @cleoabram and author @stevenbjohnson explore how NotebookLM might accelerate Cleo’s research process and second brain workflow.

https://t.co/Efpm7IDVWG

[Andrej Karpathy @karpathy]

# On the "hallucination problem"

I always struggle a bit with I'm asked about the "hallucination problem" in LLMs. Because, in some sense, hallucination is all LLMs do. They are dream machines.

We direct their dreams with prompts. The prompts start the dream, and based on the LLM's hazy recollection of its training documents, most of the time the result goes someplace useful.

It's only when the dreams go into deemed factually incorrect territory that we label it a "hallucination". It looks like a bug, but it's just the LLM doing what it always does.

At the other end of the extreme consider a search engine. It takes the prompt and just returns one of the most similar "training documents" it has in its database, verbatim. You could say that this search engine has a "creativity problem" - it will never respond with something new. An LLM is 100% dreaming and has the hallucination problem. A search engine is 0% dreaming and has the creativity problem.

All that said, I realize that what people *actually* mean is they don't want an LLM Assistant (a product like ChatGPT etc.) to hallucinate. An LLM Assistant is a lot more complex system than just the LLM itself, even if one is at the heart of it. There are many ways to mitigate hallcuinations in these systems - using Retrieval Augmented Generation (RAG) to more strongly anchor the dreams in real data through in-context learning is maybe the most common one. Disagreements between multiple samples, reflection, verification chains. Decoding uncertainty from activations. Tool use. All an active and very interesting areas of research.

TLDR I know I'm being super pedantic but the LLM has no "hallucination problem". Hallucination is not a bug, it is LLM's greatest feature. The LLM Assistant has a hallucination problem, and we should fix it.

</rant> Okay I feel much better now :)

[Marko Cadez @cadezmarko]

Check out the latest article in my newsletter: UPSIDE-DOWN PEOPLE https://t.co/Lh2Y15QR3o via @LinkedIn

[Marko Cadez @cadezmarko]

@zrh_airport you definitely managed to be the worst airport service a word. Even @JFKairport in the middle of #UNGA2023 was like VIP lounge for this. One of 5 lines for security control is open 👐

[Marko Cadez @cadezmarko]

@berlinairport if I paid and got confirmed for your Airport assistance in order to get to the meeting on time and nobody shows have I misunderstood the words Assistance and Airport? #justasking