ErikJonker, to ai
@ErikJonker@mastodon.social avatar

Ofcourse results needs to be verified and confirmed in practice but after reading the
MedGemini paper from Google there is no doubt in my mind AI will change the world of medicines. Not replacing people but augmenting them during diagnosis, operations and treatment of patients.
https://arxiv.org/abs/2404.18416

doctorambient, to LLMs
@doctorambient@mastodon.social avatar

People: stop asking to explain their behavior.

We already know that LLMs don't have the introspection necessary to explain their behavior, and their explanations are often fanciful or "just wrong."

For instance, Gemini claims it reads your emails for training, Google says it doesn't.

(BTW, if it turns out Gemini is right and Google is lying, that might be another example of an LLM convincing me it's actually "intelligent.")

reiver, to random
@reiver@mastodon.social avatar

I have been on the Fediverse long enough to remember when the hash-tag was about the (rather than or the ).

ErikJonker, to ai
@ErikJonker@mastodon.social avatar

Als je Google Gemini vergelijkt met https://perplexity.ai , dan is het onvoorstelbaar hoe slecht Gemini is, als je bedenkt hoe groot het bedrijf is wat er achter zit (Google), echt een aanrader om uit te proberen Perplexity.

ErikJonker, to ai Dutch
@ErikJonker@mastodon.social avatar

Google is at least a year behind GPT4 with the quality of it's model. At the same time OpenAI makes an absolute mess of the whole idea around customGPTs , very little customGPTs that add value. If Google starts integrating Gemini in it's ecosystem and can faster improve it's model, OpenAI is in trouble. Maybe OpenAI should let the customerfacing parts to Microsoft and focus on the model itself.

ErikJonker, to llm
@ErikJonker@mastodon.social avatar

Nice balanced article about what LLMs can and can't do, how they compare.
https://www.oneusefulthing.org/p/which-ai-should-i-use-superpowers

ErikJonker, to ai
@ErikJonker@mastodon.social avatar

Gemini Advanced is still so bad i don't think about keeping it when my free trial is over. Especially the hallucinations when asking fact-based questions are sometimes very bad, ChatGPTplus seems to have more control in that respect.

kopio, to random
@kopio@mas.to avatar

Me: (ask question)
G: It cannot be done.
Me: (select "more professional" option)
G: It can be done!

indianewswatch, to ai
@indianewswatch@kolektiva.social avatar

What do a minister’s threats against Google’s chatbot mean for AI in India?

While artificial intelligence regulation is in its infancy globally, experts think the Modi government’s belligerence will hurt the industry.

https://scroll.in/article/1064700/what-does-a-ministers-threats-against-googles-chatbot-mean-for-ai-in-india

FoW, to random Korean
@FoW@netsphere.one avatar

구글이 시간도 허헌하는 경지에 이름. 이제 Hey Google 못 쓸 듯.

TechDesk, to ai
@TechDesk@flipboard.social avatar

Google has temporarily suspended the ability for its AI chatbot Gemini to create images of people, amid a backlash over its inaccuracies. Screenshots were shared via social media of the chatbot's inability to generate images of caucasian people, seemingly showing a preference to depict people with darker skin.

Google temporarily disables AI image generation of people in Gemini:
https://flip.it/LDu8KF

Google Gemini’s ‘wokeness’ sparks debate over AI censorship:
https://flip.it/mAjM5w

For more on AI follow @the

itnewsbot, to cs
@itnewsbot@schleuss.online avatar

Google goes “open AI” with Gemma, a free, open-weights chatbot family - Enlarge (credit: Google)

On Wednesday, Google announced a new ... - https://arstechnica.com/?p=2005035 #source-availableai #machinelearning #openweightsai #googlegemini #openaimodels #metallama #chatgpt #chatgtp #biz#google #gemini #llama2 #openai #llama #meta #ai

pbeens, to ChatGPT
@pbeens@mstdn.ca avatar
pbeens,
@pbeens@mstdn.ca avatar

@ErikJonker, I have yet to investigate the Google Workspace integration with , mostly because I don't know if it's available to me yet in Canada. I sure hope that's the feature that convinces me to pay for it after the two-month trial runs out.

pbeens, to ChatGPT
@pbeens@mstdn.ca avatar

. can't make a 16x9 image, even if I tell it to. It keeps making the image square. worked okay, once I scolded it. BTW here's what ChatGPT gave me.

ErikJonker,
@ErikJonker@mastodon.social avatar

@pbeens ..i can't even make a picture with in europe...

daeinc, to machinelearning
@daeinc@genart.social avatar

instructed Google Gemini model to write this sketch. This is a result after a few back and forth chatting.

video/mp4

moribundo, to random Spanish
@moribundo@hispagatos.space avatar

protocol is not IA. Please, use the tags correctly so as not to confuse users.

daeinc, to random
@daeinc@genart.social avatar

Asking GoogleGemini to comment on creative work.

After seeing Shiffbot and GrantCuster's demos, I thought I'd make my own prototype. Nice that it can include web links in the response.

video 2x speed.

itnewsbot, to machinelearning
@itnewsbot@schleuss.online avatar

Report: Sam Altman seeking trillions for AI chip fabrication from UAE, others - Enlarge / OpenAI Chief Executive Officer Sam Altman walks on the House ... - https://arstechnica.com/?p=2002331

xekri, to esperanto

Today I tried Gemini Advanced using my Esperanto translation test question: "On Thursday, I will go to the park and walk my dog." Many other LLMs get "walk my dog" wrong. They use "promenos" which is an intransitive verb.

Let's compare:

GPT-4 (OpenAI): Ĵaŭde, mi iros al la parko kaj promenos mian hundon.

Gemini Advanced (Google): Ĵaŭde, mi iros al la parko kaj promenigos mian hundon.

Gemini (formerly Bard) (Google): Ĵaŭdon mi iros al la parko kaj promendigos mian hundon.

Claude (Anthropic): Ĝisdaten mi iros al la parko kaj promenos mian hundon.

Congrats to Google's Gemini Advanced for getting it right!

aperalesf, to random Spanish
@aperalesf@masto.es avatar

is an alternative internet protocol born in 2019. Please use to the AI-related stuff. Gemini protocol is all the oposite from AI.

ErikJonker, to ai
@ErikJonker@mastodon.social avatar

Even if the launch and performance were slightly underwhelming, the Google Gemini paper is an interesting read,
https://arxiv.org/abs/2312.11805

ErikJonker, to ai
@ErikJonker@mastodon.social avatar

Basically Google had a year to catch up with OpenAI but until now i think they have failed. Their Gemini model is not really that impressive. Ofcourse a massive worldwide rollout and availability could change things and maybe some improvements will follow or Gemini Ultra will dazzle us next year, time will tell.
https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard

chrisoffner3d, to ComputerScience
  • All
  • Subscribed
  • Moderated
  • Favorites
  • Leos
  • Durango
  • ngwrru68w68
  • thenastyranch
  • magazineikmin
  • hgfsjryuu7
  • DreamBathrooms
  • Youngstown
  • slotface
  • vwfavf
  • PowerRangers
  • everett
  • kavyap
  • rosin
  • anitta
  • khanakhh
  • tacticalgear
  • InstantRegret
  • cubers
  • mdbf
  • ethstaker
  • osvaldo12
  • GTA5RPClips
  • cisconetworking
  • tester
  • normalnudes
  • modclub
  • provamag3
  • All magazines