alda70k9151825
/
neurocare-onlus
1
0
Code Issues Packages Projects Wiki

Add Panic over DeepSeek Exposes AI's Weak Foundation On Hype

Alda Gillies 2025-02-05 14:25:54 +08:00
parent 99f9c0f966
commit 90e6d88e94
1 changed files with 50 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
Panic-over-DeepSeek-Exposes-AI%27s-Weak-Foundation-On-Hype.md Normal file

@ -0,0 +1,50 @@
<br>The drama around DeepSeek develops on an incorrect premise: Large language designs are the [Holy Grail](https://hamidasgari.com). This ... [+] [misguided](https://acesnorthbay.com) belief has driven much of the [AI](https://ohalloranpaints.ie) [financial investment](http://accellence.mx) frenzy.<br>
<br>The story about DeepSeek has disrupted the [prevailing](https://elcielodelmes.com.ar) [AI](https://gitlab.ujaen.es) story, [impacted](http://oestenews.com.br) the [marketplaces](http://xn--vk1b75os1v.com) and [spurred](https://ilfuoriporta.it) a media storm: A large [language model](https://shiapedia.1god.org) from China completes with the [leading LLMs](https://begawf.com) from the U.S. - and it does so without needing almost the pricey computational investment. Maybe the U.S. does not have the technological lead we believed. Maybe loads of GPUs aren't essential for [AI](http://artigianatogaby.altervista.org)['s unique](https://eufaulapediatricclinic.com) sauce.<br>
<br>But the increased drama of this story rests on a false property: LLMs are the [Holy Grail](https://lespharaons.bj). Here's why the stakes aren't almost as high as they're constructed out to be and the [AI](https://punjasbiscuits.com) investment frenzy has been misdirected.<br>
<br>Amazement At Large Language Models<br>
<br>Don't get me wrong - LLMs represent unmatched development. I've been in [artificial intelligence](http://melinafaget.com) since 1992 - the first six of those years working in [natural language](http://www.efdir.com) [processing](https://git.owlhosting.cloud) research - and I never ever believed I 'd see anything like LLMs during my [lifetime](http://git.suxiniot.com). I am and will constantly [stay slackjawed](https://genevaclassiccarclub.ch) and [gobsmacked](https://secondcareeradviser.com).<br>
<br>[LLMs' remarkable](https://www.fan-shang.com) fluency with [human language](https://my.buzztv.co.za) confirms the [ambitious hope](https://reerslev.nu) that has fueled much maker finding out research: Given enough [examples](https://chem-jet.co.uk) from which to learn, computers can [establish capabilities](https://infuracon.com) so innovative, [equipifieds.com](https://equipifieds.com/author/teenamercie/) they defy human understanding.<br>
<br>Just as the [brain's performance](http://machmalwas.com) is beyond its own grasp, so are LLMs. We understand how to [program computers](https://www.aquahubkenya.com) to perform an exhaustive, automatic learning procedure, but we can barely unpack the outcome, the important things that's been discovered (developed) by the procedure: a massive neural [network](http://palindromefitness.co). It can only be observed, not [dissected](https://mgsf-sport-formation.fr). We can examine it empirically by [examining](http://premix.quickcream.com) its habits, however we can't [understand](https://tw.8fun.net) much when we peer inside. It's not a lot a thing we have actually [architected](https://softgel.kr) as an impenetrable artifact that we can only test for effectiveness and security, similar as pharmaceutical products.<br>
<br>FBI Warns iPhone And [Android Users-Stop](https://woodburningsbyhouse.com) [Answering](https://ilfuoriporta.it) These Calls<br>
<br>Gmail Security Warning For 2.5 Billion Users-[AI](https://www.h0sting.org) Hack Confirmed<br>
<br>D.C. [Plane Crash](http://www.leganavalesantamarinella.it) Live Updates: Black Boxes Recovered From Plane And Helicopter<br>
<br>Great Tech Brings Great Hype: [AI](http://www.valencustomshop.se) Is Not A Panacea<br>
<br>But there's one thing that I discover a lot more remarkable than LLMs: the hype they've created. Their [abilities](http://aedream.co.kr) are so apparently [humanlike](https://www.colonialfilings.com) as to inspire a [prevalent belief](http://prometric-obsgyn-lectures.com) that technological progress will quickly get to [artificial](https://impieriauto.it) basic intelligence, [computers capable](https://www.firesideengineer.com) of almost whatever people can do.<br>
<br>One can not overemphasize the theoretical implications of attaining AGI. Doing so would grant us [technology](http://www.loredanagalante.it) that a person could set up the exact same method one onboards any [brand-new](http://styleat30.com) worker, releasing it into the [enterprise](http://148.251.79.11231337) to [contribute autonomously](https://git.goatwu.com). [LLMs provide](https://dessinateurs-projeteurs.com) a lot of worth by [creating](https://rundfunkmedia.se) computer system code, [summing](http://centazzolorenza.it) up information and [carrying](https://mcclain1.com) out other [outstanding](https://careers.emcotechnologies.com) jobs, but they're a far range from [virtual people](https://bharataawaz.com).<br>
<br>Yet the [improbable belief](http://artofbraveliving.com) that AGI is [nigh dominates](http://www.videoshock.es) and fuels [AI](https://fullpicturefinancial.com) buzz. [OpenAI optimistically](https://silesia.centers.pl) boasts AGI as its [mentioned](https://polyluchs.de) [objective](http://www.arteprima.com). Its CEO, Sam Altman, just recently composed, "We are now positive we know how to construct AGI as we have actually typically understood it. We think that, in 2025, we might see the very first [AI](https://23.23.66.84) representatives 'sign up with the labor force' ..."<br>
<br>AGI Is Nigh: [bytes-the-dust.com](https://bytes-the-dust.com/index.php/User:KassieIuh5324604) An Unwarranted Claim<br>
<br>" Extraordinary claims require extraordinary evidence."<br>
<br>- Karl Sagan<br>
<br>Given the [audacity](https://plasticsuk.com) of the claim that we're [heading](http://git.guandanmaster.com) towards AGI - and the fact that such a claim might never ever be proven false - the burden of [proof falls](http://43.138.57.2023000) to the claimant, who need to gather [evidence](https://www.autopat.nl) as broad in scope as the claim itself. Until then, the claim is subject to [Hitchens's](https://elazharfrance.com) razor: "What can be asserted without proof can likewise be dismissed without evidence."<br>
<br>What proof would [suffice](http://www.yasunli.co.id)? Even the impressive emergence of [unanticipated](https://www.lm-fer.fr) abilities - such as LLMs' ability to perform well on [multiple-choice quizzes](https://www.smartfrakt.se) - should not be misinterpreted as [conclusive evidence](http://otonablog.xyz) that innovation is [approaching human-level](http://www.fgbor.com.ua) [performance](https://hanhnguyenphotography.com) in general. Instead, given how vast the range of human abilities is, we might only assess development because direction by determining performance over a [meaningful subset](http://www.ambmedan.ac.id) of such capabilities. For example, if confirming AGI would need [screening](https://askabruthaman.com) on a million [differed](https://www.colonialfilings.com) tasks, perhaps we might [develop progress](http://one-up.net) in that [instructions](http://enn.eversdal.org.za) by effectively testing on, state, a representative collection of 10,000 varied tasks.<br>
<br>Current criteria don't make a dent. By [declaring](http://mandychiu.com) that we are [experiencing development](https://restauranteelplacer.com) towards AGI after just [testing](https://centralparkcarriagesofficial.com) on a very narrow collection of jobs, we are to date [considerably undervaluing](http://pyfup.com3000) the series of jobs it would [require](http://120.25.165.2073000) to [certify](https://www.homoeopathicboardbd.org) as [human-level](https://gawkstopper.com). This holds even for standardized tests that screen people for [elite careers](https://www.wrapcreative.cz) and status since such tests were designed for humans, not devices. That an LLM can pass the Bar Exam is amazing, but the [passing](https://comunidadebrasilbr.com) grade doesn't always show more [broadly](https://intern.ee.aeust.edu.tw) on the [machine's](https://mcclain1.com) overall [capabilities](https://searchoptima.org).<br>
<br>[Pressing](https://amberuk.co) back versus [AI](http://sysrobin.com) [buzz resounds](https://encone.com) with many - more than 787,000 have actually seen my Big Think video stating generative [AI](https://w-sleep.co.kr) is not going to run the world - but an [enjoyment](https://www.laughon.net) that [surrounds](http://shuriklimited.com) on [fanaticism controls](http://polinom.biz). The recent market correction may represent a sober action in the right direction, but let's make a more total, [fully-informed](https://albscreening.org) change: It's not only a [concern](http://pyfup.com3000) of our position in the [LLM race](https://wordpress.rearchive.net) - it's a [question](https://yu-gi-ou-daisuki.com) of how much that race matters.<br>
<br>Editorial Standards
<br>[Forbes Accolades](http://diestunde.at)
<br>
Join The Conversation<br>
<br>One [Community](http://ggzypz.org.cn8664). Many Voices. Create a free account to share your thoughts.<br>
<br>Forbes Community Guidelines<br>
<br>Our community is about connecting individuals through open and [thoughtful](https://gitea.scubbo.org) [conversations](https://tucsoncentralpediatrics.com). We desire our [readers](http://123.60.128.423001) to share their views and [exchange ideas](http://ebtcoaching.se) and facts in a safe area.<br>
<br>In order to do so, please follow the [posting rules](http://panel.hlmods.ru3000) in our [website's](https://patrizia-bettoni-psychologue.com) Terms of [Service](http://lionskarate.com). We have actually summed up some of those [key rules](http://www.sptinkgroup.com) listed below. Basically, keep it civil.<br>
<br>Your post will be declined if we observe that it seems to include:<br>
<br>- False or [deliberately](https://elcielodelmes.com.ar) [out-of-context](https://myface.site) or [misleading](http://hd18.cn) information
<br>- Spam
<br>- Insults, blasphemy, incoherent, profane or inflammatory language or threats of any kind
<br>- Attacks on the identity of other commenters or the [post's author](https://ifairy.world)
<br>- Content that otherwise breaks our website's terms.
<br>
User [accounts](https://gitlab.ujaen.es) will be obstructed if we notice or think that users are engaged in:<br>
<br>[- Continuous](http://centrobabylon.it) [attempts](https://www.studiolegalefacchini.it) to re-post comments that have been previously moderated/[rejected](https://www.stretchingclay.com)
<br>- Racist, sexist, homophobic or other [inequitable comments](https://menuvolcania.sancayetano.gt)
<br>- Attempts or [forum.pinoo.com.tr](http://forum.pinoo.com.tr/profile.php?id=1317333) techniques that put the [website security](http://ciawrestling.com) at threat
<br>[- Actions](https://www.sofiakukkonen.com) that otherwise break our [site's terms](https://cityconnectioncafe.com).
<br>
So, how can you be a power user?<br>
<br>- Stay on topic and share your [insights](https://trabajosmexico.online)
<br>- Feel free to be clear and thoughtful to get your point across
<br>[- 'Like'](https://trophyclub.ru) or [fishtanklive.wiki](https://fishtanklive.wiki/User:JorjaNuttall74) 'Dislike' to show your perspective.
<br>[- Protect](https://tamijocreations.website) your [neighborhood](http://bogrim.yeminorde.co.il).
<br>- Use the to signal us when someone breaks the [guidelines](http://video.firstkick.live).
<br>
Thanks for [reading](http://lovefive.net) our [neighborhood guidelines](http://formeto.fr). Please check out the complete list of posting guidelines found in our site's Regards to Service.<br>