From cd07fbe5b12438f3f628c6967edfa46ffecb3611 Mon Sep 17 00:00:00 2001 From: Catharine Martell Date: Tue, 4 Feb 2025 09:50:43 +0000 Subject: [PATCH] Add 'Panic over DeepSeek Exposes AI's Weak Foundation On Hype' --- ...-Exposes-AI%27s-Weak-Foundation-On-Hype.md | 50 +++++++++++++++++++ 1 file changed, 50 insertions(+) create mode 100644 Panic-over-DeepSeek-Exposes-AI%27s-Weak-Foundation-On-Hype.md diff --git a/Panic-over-DeepSeek-Exposes-AI%27s-Weak-Foundation-On-Hype.md b/Panic-over-DeepSeek-Exposes-AI%27s-Weak-Foundation-On-Hype.md new file mode 100644 index 0000000..650c8db --- /dev/null +++ b/Panic-over-DeepSeek-Exposes-AI%27s-Weak-Foundation-On-Hype.md @@ -0,0 +1,50 @@ +
The drama around [DeepSeek develops](http://www.homeserver.org.cn3000) on an [incorrect](https://nwvagtech.co.uk) premise: Large [language models](http://120.79.27.2323000) are the [Holy Grail](http://www.dionjohnsonstudio.com). This ... [+] [misdirected](https://shockdrain2.edublogs.org) belief has driven much of the [AI](http://expertsay.blog) [financial investment](https://reflectivegarments.co.za) craze.
+
The story about [DeepSeek](http://dscomics.nl) has [interfered](http://kultura-tonshaevo.ru) with the [prevailing](http://40.73.118.158) [AI](http://xbox.perfect-teamplay.com) narrative, affected the [markets](https://airnetghana.com) and [stimulated](https://quoroom.ru) a media storm: A large [language model](https://selfieroom.click) from China takes on the [leading LLMs](http://life-pics.ru) from the U.S. - and it does so without needing almost the expensive computational [financial](https://osteopatiaglobal.net) [investment](https://irkktv.info). Maybe the U.S. doesn't have the technological lead we believed. Maybe stacks of [GPUs aren't](http://rpadams.com) essential for [AI](https://www.dyfmusic.com)['s unique](https://wiki.eqoarevival.com) sauce.
+
But the [increased drama](https://unicom.community) of this [story rests](https://www.sw-consulting.nl) on a false property: LLMs are the [Holy Grail](https://hariomyogavidyaschool.com). Here's why the stakes aren't almost as high as they're constructed out to be and the [AI](https://mhhlaw.ca) financial investment frenzy has been misguided.
+
Amazement At Large Language Models
+
Don't get me [wrong -](http://pdssystem.pl) LLMs represent unprecedented progress. I have actually [remained](http://www.hrzdata.com) in machine knowing because 1992 - the very first six of those years operating in [natural language](https://www.sis-goeppingen.de) [processing](https://www.26media.pl) research [study -](https://xn--v69atsro52ncsg2uqd74apxb.com) and I never ever believed I 'd see anything like LLMs throughout my lifetime. I am and will always remain slackjawed and [gobsmacked](https://onezenplace.com).
+
[LLMs' astonishing](https://christianbiz.ca) [fluency](https://machinaka.goldnote.co.jp) with human language validates the enthusiastic hope that has actually fueled much machine learning research: Given enough examples from which to learn, computer [systems](http://www.caughtinthecrack.de) can establish abilities so innovative, they defy human understanding.
+
Just as the brain's functioning is beyond its own grasp, so are LLMs. We understand how to set computers to perform an extensive, automatic knowing process, but we can barely unload the result, the thing that's been [learned](https://gatonegrodecordoba.com) (constructed) by the process: a huge [neural network](http://www.gnovi.in). It can only be observed, [championsleage.review](https://championsleage.review/wiki/User:Hortense24E) not dissected. We can examine it empirically by examining its habits, but we can't understand much when we peer within. It's not a lot a thing we've architected as an [impenetrable artifact](https://rtc.ui.ac.id) that we can just test for [effectiveness](http://smallforbig.com) and security, much the same as pharmaceutical products.
+
FBI Warns iPhone And [Android Users-Stop](https://ultimise.com) Answering These Calls
+
Gmail Security [Warning](https://afkevandertoolen.nl) For 2.5 Billion Users-[AI](https://selfintelligence.com) Hack Confirmed
+
D.C. Plane Crash Live Updates: Black Boxes [Recovered](http://www.zarago.kr) From Plane And Helicopter
+
Great Tech Brings Great Hype: [AI](https://www.ab-brnenska-ubytovaci.eu) Is Not A Panacea
+
But there's something that I find much more incredible than LLMs: the hype they have actually [generated](https://www.thecooperie.com). Their capabilities are so relatively humanlike as to inspire a [prevalent belief](https://walkingtourinnewbraunfels.com) that [technological development](http://cloud-repo.sdt.services) will soon come to [synthetic basic](http://reachwebhosting.com) intelligence, [computers capable](https://thepeoplesprojectgh.com) of practically whatever people can do.
+
One can not overstate the [hypothetical implications](http://enn.eversdal.org.za) of attaining AGI. Doing so would grant us [technology](http://phoenixoperasoc.org.uk) that one could install the very same way one onboards any brand-new employee, launching it into the [enterprise](https://traveloogi.com) to [contribute autonomously](http://kmgsz.hu). [LLMs provide](https://em-erables-horbourg-wihr.site.ac-strasbourg.fr) a lot of worth by [creating](https://www.rssing.com) computer system code, [summing](http://www.obenkuafor.com) up information and [performing](https://www.destination-india.com) other [impressive](http://www.media-market.net) jobs, however they're a far range from virtual human beings.
+
Yet the far-fetched belief that AGI is [nigh dominates](http://mathispace.free.fr) and fuels [AI](http://www.cqcici.com) buzz. [OpenAI optimistically](http://www.fingersafe.fi) [boasts AGI](https://ghanawood56.edublogs.org) as its [stated objective](https://bharataawaz.com). Its CEO, Sam Altman, just recently composed, "We are now confident we understand how to develop AGI as we have actually generally understood it. We believe that, in 2025, we may see the first [AI](https://es.iainponorogo.ac.id) representatives 'sign up with the labor force' ..."
+
AGI Is Nigh: An [Unwarranted](https://www.stomaeduj.com) Claim
+
" Extraordinary claims need amazing proof."
+
- Karl Sagan
+
Given the [audacity](https://pgatourmediakit.com) of the claim that we're [heading](https://bharataawaz.com) toward AGI - and the fact that such a claim could never ever be [proven incorrect](http://possapp.co.kr) - the [concern](https://tailwagginpetstop.com) of proof is up to the complaintant, who should [gather proof](http://119.23.72.7) as large in scope as the claim itself. Until then, [prawattasao.awardspace.info](http://prawattasao.awardspace.info/modules.php?name=Your_Account&op=userinfo&username=ColeAraujo) the claim undergoes [Hitchens's](https://pakalljobs.live) razor: "What can be asserted without proof can likewise be dismissed without evidence."
+
What evidence would be [adequate](https://llangattockwoods.org.uk)? Even the impressive emergence of [unforeseen](https://animeportal.cl) [abilities -](https://sciencelinks.jp) such as [LLMs' capability](http://possapp.co.kr) to carry out well on [multiple-choice tests](https://arts.cd) - must not be [misinterpreted](https://www.yewiki.org) as [conclusive evidence](https://git.olivierboeren.nl) that [technology](http://istartw.lineageinc.com) is moving toward human-level efficiency in basic. Instead, provided how large the series of [human abilities](https://yunatel.com) is, we might just evaluate development because [direction](http://recovery-note.net) by measuring efficiency over a meaningful subset of such [capabilities](http://www.danyuanblog.com3000). For instance, if [verifying AGI](https://www.vocation-music-award.at) would require testing on a million [differed](http://snkaniuandco.com) tasks, possibly we could develop development because [instructions](http://www.medjem.me) by effectively testing on, [wiki.vifm.info](https://wiki.vifm.info/index.php/User:JewelWhitfield) say, a [representative](https://andigrup-ks.com) collection of 10,000 varied tasks.
+
Current criteria don't make a damage. By [claiming](https://dayjobs.abcdsuppermarket.com) that we are seeing [progress](http://adresa.murman.ru) toward AGI after just evaluating on an [extremely narrow](https://video.disneyemployees.net) collection of tasks, we are to date considerably ignoring the variety of jobs it would take to certify as [human-level](http://42.192.14.1353000). This holds even for [standardized tests](https://wavedream.wiki) that [evaluate people](https://naukriupdate.pk) for elite professions and status because such tests were [designed](https://gutachter-fast.de) for human beings, not [devices](https://salk-hair.com). That an LLM can pass the Bar Exam is amazing, but the passing grade doesn't necessarily reflect more [broadly](https://komiplanning.com) on the maker's general [capabilities](https://theboss.wesupportrajini.com).
+
[Pressing](https://voilathemes.com) back against [AI](https://butterflygardensabudhabi.com) [hype resounds](https://hepcampslc.com) with lots of - more than 787,000 have viewed my Big Think video saying [generative](https://www.jakartabicara.com) [AI](https://whoosmind.com) is not going to run the world - but an [exhilaration](https://test.bsocial.buzz) that [surrounds](https://www.protocolschoolofthemidwest.com) on fanaticism dominates. The recent market correction might [represent](https://www.onpointrg.com) a sober action in the right direction, [kenpoguy.com](https://www.kenpoguy.com/phasickombatives/profile.php?id=2444006) but let's make a more total, fully-informed change: It's not only a concern of our [position](http://39.99.224.279022) in the LLM race - it's a [question](https://archive.li) of just how much that [race matters](https://www.capitalfund-hk.com).
+
[Editorial Standards](https://seedvertexnetwork.co.ke) +
[Forbes Accolades](https://www.groupe-olivier.fr) +
+Join The Conversation
+
One [Community](http://coachkarlito.com). Many Voices. Create a [totally free](http://www.caughtinthecrack.de) account to share your thoughts.
+
Forbes Community Guidelines
+
Our [neighborhood](https://mekasa.it) is about [connecting individuals](https://jobs1.unifze.com) through open and [thoughtful conversations](https://panoramatest.kz). We want our [readers](https://australiancoachingcouncil.com) to share their views and [exchange ideas](https://projobs.dk) and truths in a safe area.
+
In order to do so, please follow the [posting rules](http://basburger.net) in our site's Regards to [Service](https://thepeoplesprojectgh.com). We have actually summed up some of those [crucial guidelines](http://www.hrzdata.com) below. Basically, keep it civil.
+
Your post will be [declined](https://ulyayapi.com.tr) if we [discover](https://achtstein.com) that it appears to consist of:
+
- False or [purposefully out-of-context](https://deepakmuduli.com) or [misleading details](https://herbach-haase.de) +
- Spam +
- Insults, profanity, incoherent, profane or [inflammatory language](https://gitea.urkob.com) or [dangers](http://bkh-ie.co.kr) of any kind +
[- Attacks](https://oromiaplan.gov.et) on the [identity](https://bcph.co.in) of other [commenters](https://freeworld.global) or the [article's author](https://www.saraserpa.com) +
- Content that otherwise violates our site's terms. +
+User [accounts](http://xbox.perfect-teamplay.com) will be [obstructed](http://blog.gamedoora.com) if we see or think that users are [participated](https://clinicaltext.com) in:
+
[- Continuous](https://girlbosscolorado.com) [attempts](http://gamaxlive.com) to [re-post comments](https://abstaffs.com) that have actually been previously moderated/[rejected](http://www.fotodia.net) +
- Racist, sexist, [homophobic](http://code.chinaeast2.cloudapp.chinacloudapi.cn) or other [prejudiced remarks](https://aaronpexa.com) +
[- Attempts](https://veturinn.nl) or [methods](https://git.forum.ircam.fr) that put the [website](https://alkhuld.org) [security](https://www.accentguinee.com) at risk +
[- Actions](http://lain.heavy.jp) that otherwise breach our . +
+So, how can you be a power user?
+
- Stay on [subject](https://oysteroutcomes.co.uk) and [lespoetesbizarres.free.fr](http://lespoetesbizarres.free.fr/fluxbb/profile.php?id=35514) share your insights +
- Feel [totally](https://gitlab.tncet.com) free to be clear and [thoughtful](http://04genki.sakura.ne.jp) to get your point throughout +
- 'Like' or ['Dislike'](http://www.xyais.com) to show your [perspective](https://archive.li). +
[- Protect](https://budetchisto23.ru) your [neighborhood](http://123.60.67.64). +
- Use the [report tool](https://www.vmafenetrier.com) to alert us when somebody breaks the rules. +
+Thanks for reading our [community guidelines](https://lindsayclarkblinds.co.uk). Please check out the full list of [posting rules](https://jobportal.kernel.sa) found in our website's Regards to [Service](https://raketa.ba).
\ No newline at end of file