Loo kontoLogi sisse

Viimatised otsingud

Pole viimatisi otsinguid

Otsimisvalikud

Saadaval vaid kui sisse logitud.
est.social on üks paljudest sõltumatutest Mastodoni serveritest, mida saab fediversumis osalemiseks kasutada.
est.social on mõeldud Eestis üldkasutatavaks Mastodoni serveriks. est.social is meant to be a general use Mastodon server for Estonia.

Administraator:

Serveri statistika:

90
aktiivsed kasutajad

est.social: TeaveProfiilikataloogIsikuandmete kaitse

Mastodon: TeaveTõmba äppKiirklahvidLähtekoodv4.3.6+est

#intelligence

47 postitusega19 osalejaga8 postitust täna
Avalik *
Karsten Johansson

People have low expectations for AI. I asked ChatGPT (paid account, not the free stuff) what would make ChatGPT better. I think the answer is a lot higher quality than what people keep saying one should expect. This is what I got:

--- My Prompt:
With DeepSeek, there are things it is very good at, but it fails quickly and rather spectacularly. Your models tend to break down as well, but somewhere there must be a model that will be able to stay on topic, even when the topic changes, so that conversations move around like human conversations move around subject-to-subject. What would take ChatGPT over the top?

--- ChatGPT reply:
Here’s what it’ll take to get closer to that human-like conversation flow:

  1. Long-Term Memory That Actually Works. Most models either have no memory or they remember only in a clunky, artificial way.

A better memory system would not just recall facts, but contextual relevance—like “we talked about LLM efficiency two days ago and you hate cheerleader language.”

It has to be smart recall, not just data storage.

  1. Thread Awareness
    Imagine a model that can say: “You mentioned DeepSeek earlier. Should I compare it again now that we’ve brought up open-source tools?”
    That requires models to track not just words, but conversational branches—like a tree, not a straight line.

  2. Graceful Fallbacks
    When a model doesn’t know, it should gracefully pivot, not hallucinate or go off-topic. This kind of conversational “stability” is rare and subtle.

Most models that fail “spectacularly,” as you said, are trying to bluff through complexity instead of narrowing scope or clarifying.

  1. Hybrid Contextual ReasoningThe future model needs to reason across subjects rather than switching on and off.

If we're talking about language models, and you shift to science fiction, it should be able to bring up Asimov or neural networks in storytelling—bridging topics, not just hopping between them.

Right now, most models either:

Stay sharp but shallow (good at Q&A, bad at memory), or

Drift into friendly nonsense (chatty, but brittle)

The sweet spot will be a system that grows with you across time and topics—and doesn’t forget what you value in a conversation.

....

When you see stuff posted online meant to minimize the quality of output possible, This was a single question with no leader. I think this is the kind of answer we want from our AI devices.

Thoughts? Feel free to tear this response apart. It needs that kind of human analysis.

#ai#openai#deepseek
Avalik *
J.P. Wing

Apple iOS users, I'm curious. Are you using Apple Intelligence in any way on your iPhone? Boosts appreciated for the poll.

· 16 inimest ·
#apple#intelligence#ios
Avalik
Alexandre Dulaunoy

A new release of the AIL project is coming soon, featuring a significant improvement in language detection.

A lot of work has been done on LexiLang by @terrtia to clean up dictionaries and improve support for localized languages and slang.

In the example below, you can see a user active in different Telegram channels, using both Russian and Ukrainian.

🔗 ail-project.org/

If you're interested in the topic, join us at a 2-day hackathon in Luxembourg on April 8–9, 2025, focused on open-source security tools. The developers of the AIL project will be there in person!

🔗 hackathon.lu/

#threatintel #threatintelligence #opensource #ail #intelligence

@ail_project
@circl

Language detection in the next release of AIL project. We can see that a user on different telegram channels used different languages.
Jätkatud lõim
Avalik
Nonilex

#Arctic #warming could pose a threat to America’s #security too: #Alaska may have new vulnerabilities to both #China & #Russia; changes in #ocean salinity & temp might interfere w/ #submarine detection systems; extremes of climate change, including #permafrost thaw in Russia, may drive #economic #instability, social #unrest, & territorial claims. [#Trump #Intelligence Agencies excluded #ClimateChange from the annual #ThreatAssessment report for the first time in decades — too *woke*]

#Greenland
Avalik
Nonilex

#Trump’s instructions to his team late this week remained to attack #JeffreyGoldberg & #TheAtlantic, & not “produce a scalp” by firing any members of his Cabinet, in the words of one outside adviser consulted by the White House who described the discussions to us.

#law #NationalSecurity #military #intelligence #SignalGate #Signal #idiocracy #kakistocracy #propaganda #authoritarianism #autocracy #tyranny
theatlantic.com/politics/archi

The Atlantic · Trump Resists Giving Up a 'Scalp' in SignalgateIsaac Stanley-Becker poolt
Jätkatud lõim
Avalik
Nonilex

…#Hegseth & other snr officials have argued that the information he shared was not #classified, & that it was ultimately up to his dept to decide what of its info should be considered classified.

#NationalSecurity experts said that argument fails a test of basic common sense about the danger of letting #military plans leak before an operation, & that it goes against longtime practice among military & #intelligence agencies.

#law#Trump#DualState
Avalik
Nonilex

Today’s summary (feeling done)

Other than the colossal #NationalSecurity breach by the whole of #Trump’s #intelligence & #defense cabinet secretaries (yeah…, other than than Mrs. Lincoln how was the play?) today’s news:

#TheAtlantic / #JeffreyGoldberg releases everything:

masto.ai/@Nonilex/114229022768

#Gabbard & #Ratcliffe lied through their teeth in the #House

masto.ai/@Nonilex/114229262055

Pete Hegseth: VP: I fully share your loathing of European free-loading. It's PATHETIC. But Mike is correct, we are the only ones on the planet (on our side of the ledger) who can do this. Nobody else even close. Question is timing. I feel like now is as good a time as any, given POTUS directive to reopen shipping lanes. I think we should go; but POTUS still retains 24 hours of decision space. 8:49 AM SM: As I heard it, the president was clear: green light, but we soon make clear to Egypt and Europe what we expect in return. We also need to figure out how to enforce such a requirement. EG, if Europe doesn't remunerate, then what? If the US successfully restores freedom of navigation at great cost there needs to be some further economic gain extracted in return. 9:35 AM Pete Hegseth: Agree 9:46 AM
MastodonNonilex (@Nonilex@masto.ai)Manused: 4 pilti 🚨🚨🚨And here are screenshots of the OBVIOUSLY CLASSIFIED INFORMATION texts: #law #NationalSecurity #SignalGate #Signal #idiocracy #kakistocracy #intelligence #USIntelligenceAgencies #JDVance #Hegseth #TulsiGabbard #Ratcliffe #MikeWaltz #MarcoRubio #SusieWiles #Witkoff #StephenMiller
Avalik
ahimsa

" 'Heads are exploding': How security experts see the Signal war-plan breach"

opb.org/article/2025/03/26/sig

“Any one of us that served in the military or serve in the military would be in Leavenworth if we did this"

- Former congress representative Adam Kinzinger, Republican

OPB · «Heads are exploding»: How security experts see the Signal war-plan breachBill Chappell poolt
#USPol#Intelligence#NationalSecurity
AvastaLaiv lõimed
Teave