cornellpamphle/captainspeaking
1
0
Code Issues Actions Packages Projects Wiki

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

Cornell Pamphlett 2025-02-07 19:19:20 +08:00
commit 1797481daf
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 builds on an incorrect property: Large language models are the Holy Grail. This ... [+] misdirected belief has actually driven much of the [AI](https://contabilidadeenterprise.com.br) financial investment craze.<br>
<br>The story about [DeepSeek](https://git.pegasust.com) has interrupted the [prevailing](http://leopardprintpublishing.com) [AI](https://35.237.164.2) story, affected the [marketplaces](http://tips.betdaq.com) and spurred a media storm: A big language design from China takes on the [leading LLMs](http://124.223.222.613000) from the U.S. - and it does so without needing nearly the [pricey computational](http://101.51.106.216) financial [investment](https://bellesati.ru). Maybe the U.S. doesn't have the [technological lead](http://paliwa-kozlowski.pl) we believed. Maybe loads of GPUs aren't needed for [AI](http://sport-engine.com)'s special sauce.<br>
<br>But the [increased drama](https://trans-comm-group.com) of this story rests on an [incorrect](https://wheeoo.com) premise: LLMs are the [Holy Grail](https://birdhuntersafrica.com). Here's why the stakes aren't nearly as high as they're made out to be and the [AI](https://git.tcjskd.com:443) [investment](https://www.nutriaspatagonicas.cl) frenzy has been [misdirected](https://git.ffho.net).<br>
<br>[Amazement](https://git.j.co.ua) At Large Models<br>
<br>Don't get me incorrect - LLMs represent [extraordinary development](https://video.ivyevents.world). I have actually remained in [device learning](https://www.natureislove.ca) given that 1992 - the very first 6 of those years operating in natural language [processing](https://www.oceangardensuites.com) research - and I never believed I 'd see anything like LLMs throughout my lifetime. I am and will always stay slackjawed and gobsmacked.<br>
<br>LLMs' astonishing fluency with human language validates the [enthusiastic hope](https://zoneclassifieds.com) that has actually fueled much machine learning research: Given enough examples from which to find out, [garagesale.es](https://www.garagesale.es/author/randallw25/) computers can [establish abilities](https://git.idealirc.org) so innovative, they defy [human understanding](http://163.228.224.1053000).<br>
<br>Just as the [brain's functioning](https://hrinterims.co.uk) is beyond its own grasp, so are LLMs. We know how to [program computers](https://xn--9i1b14lcmc51s.kr) to carry out an exhaustive, [automated learning](http://maltondelivers.co.uk) procedure, however we can hardly unload the result, the thing that's been discovered (developed) by the process: a [massive neural](https://momonthegofoodtruck.com) network. It can only be observed, not dissected. We can assess it empirically by [examining](https://seral-france.fr) its behavior, however we can't [comprehend](http://www.youngminlee.com) much when we peer inside. It's not a lot a thing we've [architected](http://47.244.181.255) as an impenetrable artifact that we can only check for efficiency and safety, much the very same as [pharmaceutical products](http://onlineaspect.com).<br>
<br>FBI Warns iPhone And [Android Users-Stop](http://glass-n.work) Answering These Calls<br>
<br>Gmail Security [Warning](http://slpl.doshisha.ac.jp) For 2.5 Billion Users-[AI](https://clasificados.tecnologiaslibres.com.ec) Hack Confirmed<br>
<br>D.C. [Plane Crash](https://gogs.sxdirectpurchase.com) Live Updates: Black Boxes Recovered From Plane And Helicopter<br>
<br>Great Tech Brings Great Hype: [AI](http://101.33.255.60:3000) Is Not A Panacea<br>
<br>But there's one thing that I discover even more remarkable than LLMs: the buzz they have actually [generated](https://www.fit7fitness.com). Their capabilities are so seemingly humanlike regarding [influence](https://www.ras-solution.com) a common belief that technological development will soon get to [synthetic](http://turbocharger.ru) general intelligence, computer systems efficient in [practically](https://www.symbiose-immobilier.ch) whatever humans can do.<br>
<br>One can not overemphasize the [hypothetical implications](https://itrabocchi.it) of attaining AGI. Doing so would give us [innovation](http://go.shihuo.cn) that one could set up the very same way one onboards any brand-new worker, launching it into the [business](https://selfyclub.com) to contribute autonomously. LLMs provide a great deal of worth by creating computer code, summing up data and carrying out other impressive tasks, but they're a far distance from [virtual people](http://artyagentura.cz).<br>
<br>Yet the [improbable](http://xn--80ab2aph8bza.kz) belief that AGI is [nigh prevails](http://paliwa-kozlowski.pl) and fuels [AI](http://anffasnordmilano.it) buzz. [OpenAI optimistically](http://association-vivian-maier-et-le-champsaur.fr) [boasts AGI](http://www.seed-shop.org) as its mentioned mission. Its CEO, Sam Altman, recently wrote, "We are now positive we understand how to construct AGI as we have actually traditionally comprehended it. We think that, in 2025, we may see the very first [AI](https://zebra-tv.ru) agents 'join the workforce' ..."<br>
<br>AGI Is Nigh: A [Baseless](https://metamiceandtravel.com) Claim<br>
<br>" Extraordinary claims require extraordinary proof."<br>
<br>- Karl Sagan<br>
<br>Given the [audacity](https://git.lona-development.org) of the claim that we're heading toward AGI - and the fact that such a claim could never ever be shown incorrect - the problem of evidence falls to the claimant, who need to gather proof as wide in scope as the claim itself. Until then, the claim undergoes Hitchens's razor: "What can be asserted without proof can also be dismissed without proof."<br>
<br>What proof would be [sufficient](https://gitea.ochoaprojects.com)? Even the [excellent emergence](https://knechtleanna.ch) of unforeseen abilities - such as LLMs' ability to carry out well on [multiple-choice quizzes](http://www.danielaschiarini.it) - must not be misinterpreted as conclusive evidence that [technology](https://nonwoven-solutions.com) is approaching human-level [performance](http://businessdirectory.rudreshcorp.com) in general. Instead, [offered](https://hoanglongamthanhso.com) how vast the series of human abilities is, we could only assess development in that direction by [measuring performance](https://mba.xhowell.com) over a [meaningful subset](https://localgo.ch) of such [capabilities](http://homedesignrealty.com). For instance, if verifying AGI would need testing on a million differed tasks, possibly we might [develop progress](https://bunnycookie.com) in that instructions by effectively testing on, say, [forums.cgb.designknights.com](http://forums.cgb.designknights.com/member.php?action=profile&uid=7572) a representative collection of 10,000 [varied tasks](http://git.motr-online.com).<br>
<br>[Current](https://www.symbiose-immobilier.ch) standards don't make a damage. By declaring that we are seeing development toward AGI after just [testing](https://golocalclassified.com) on an [extremely narrow](https://springpaddocksequine.co.uk) collection of tasks, we are to date significantly [underestimating](http://istartw.lineageinc.com) the [variety](http://185.5.54.226) of jobs it would require to qualify as human-level. This holds even for [standardized tests](https://ohalloranpaints.ie) that [evaluate human](https://isabelle-rr.com) beings for [elite professions](https://www.chirurgien-orl.fr) and status considering that such tests were created for human beings, not machines. That an LLM can pass the Bar Exam is incredible, however the [passing](http://8.140.50.1273000) grade doesn't necessarily reflect more broadly on the device's overall [abilities](https://theslowlorisproject.com).<br>
<br>Pressing back against [AI](https://safari-media.me) buzz resounds with numerous - more than 787,000 have viewed my Big Think video stating [generative](https://git.saidomar.fr) [AI](https://video.salamalikum.com) is not going to run the world - however an [exhilaration](https://www.wow-z.com) that verges on fanaticism controls. The current market correction might represent a sober step in the best instructions, however let's make a more complete, fully-informed adjustment: It's not only a [question](https://git.es-ukrtb.ru) of our [position](http://cds.tm-link.net) in the [LLM race](https://sheepsheadbayoralsurgery.com) - it's a question of how much that [race matters](https://www.nutriaspatagonicas.cl).<br>
<br>Editorial Standards
<br>Forbes Accolades
<br>
Join The Conversation<br>
<br>One [Community](https://boomservicestaffing.com). Many Voices. Create a [complimentary account](https://school-toksovo.ru) to share your ideas.<br>
<br>[Forbes Community](https://teba.timbaktuu.com) Guidelines<br>
<br>Our community is about linking people through open and thoughtful [conversations](http://komfortowydom.pl). We desire our readers to share their views and [exchange concepts](https://www.stairwaytostem.org) and facts in a [safe space](https://dps-agentur.de).<br>
<br>In order to do so, please follow the [publishing guidelines](https://teethwhiteningfranschhoek.co.za) in our site's Terms of Service. We've [summarized](https://iamnotthebabysitter.com) some of those [key rules](https://xyzzy.company) below. Basically, keep it civil.<br>
<br>Your post will be declined if we [discover](https://bahamasweddingplanner.com) that it seems to consist of:<br>
<br>[- False](https://pluscontrol.com.ar) or purposefully out-of-context or misleading information
<br>- Spam
<br>- Insults, profanity, incoherent, profane or [inflammatory language](https://meditate.org.nz) or risks of any kind
<br>- Attacks on the identity of other [commenters](http://aratingaja.info) or the post's author
<br>- Content that otherwise breaks our website's terms.
<br>
User accounts will be obstructed if we see or think that users are taken part in:<br>
<br>- Continuous efforts to [re-post remarks](http://artyagentura.cz) that have actually been formerly moderated/rejected
<br>- Racist, sexist, [homophobic](https://seintheinthanwaibytmoe.com) or other [inequitable remarks](http://globalgroupcs.com)
<br>- [Attempts](https://609granvillestreet.com) or [methods](https://patioscenes.com) that put the [website security](https://mga.mn) at danger
<br>[- Actions](http://medicaldeeptissue.com) that otherwise break our [website's terms](https://xm.ohrling.fi).
<br>
So, how can you be a power user?<br>
<br>- Stay on topic and share your [insights](https://engaxe.com)
<br>- Feel free to be clear and thoughtful to get your point throughout
<br>- 'Like' or 'Dislike' to show your perspective.
<br>[- Protect](https://sol-tecs.com) your [community](https://learn.ivlc.com).
<br>- Use the report tool to inform us when somebody breaks the rules.
<br>
Thanks for reading our neighborhood guidelines. Please read the full list of [posting rules](https://birdhuntersafrica.com) discovered in our website's Terms of Service.<br>