frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Cerebras: The AI Hardware Giant Facing Imminent Collapse

6•noob_hardy•1h ago
Post by ex-employee.

In the cutthroat AI world, Cerebras Systems, despite colossal wafers and bold claims, faces collapse. Its fragile foundation stems from brittle hardware, astonishingly narrow in practical application. Optimized primarily for LLaMA model finetuning, touted "speed boosts" often use speculative decoding, masking that many "supported" models run inefficiently, if at all. This rigidity mirrors a Formula 1 engine demanding one fuel under lab conditions.

This compounds with a poorly conceived software and kernel stack. The current infrastructure restricts Cerebras systems almost exclusively to LLaMA-family transformer LLMs. Modern vision pipelines, diffusion models, or other AI paradigms are beyond reach. In a rapidly evolving field, this narrow focus is a death sentence. A complete software overhaul is desperately needed, yet internal motivation is absent, hinting at leadership either oblivious or unwilling to invest in survival. Technical woes are exacerbated by third-class ML expertise and rudderless, top-down leadership. Non-technical decision-makers lead to misguided strategies and poor adaptation. ML expertise is subpar and highly politicized; questionable ideas from figures like Head of AI Ganesh Venkatesh are reportedly prioritized, wasting resources. An extreme resource shortage – Cerebras hardware and even GPUs – cripples meaningful work, making numbers in reports untrustworthy fiction. The chasm between rhetoric and reality widens.

A palpable resignation within the ranks further indicts the company. Employees are "waiting it out" or trading stocks, not contributing. This disengagement highlights deep problems, where personal finance trumps innovation.

Cerebras appears a bloated, incompetent, and culturally bankrupt company. Any stock price would be a scam. The rare "A players" have left or are leaving.

The writing is on the wall. Cerebras, despite grand pronouncements, is a company built on sand, destined to crumble under its limitations and poor choices.

Get out while you still can.

Why "AI Slop" Might Be the Best Thing to Happen to Visual Storytelling

https://ispo.ai/blog/ai-slop
1•olegakbarov•55s ago•1 comments

WebX11 – Turn any executable into a resizable web page

https://github.com/lp1dev/WebX11
1•leakr•2m ago•1 comments

Richard Sutton – Father of RL thinks LLMs are a dead end [video]

https://www.youtube.com/watch?v=21EYKqUsPfg
1•jonbaer•2m ago•0 comments

Did Nvidia Just Repeat Cisco's Mistake and Build a House of Cards

https://www.fool.com/investing/2025/09/28/is-nvidia-building-a-house-of-cards-with-openai-in/
1•1vuio0pswjnm7•3m ago•1 comments

OpenAI launch of video app Sora plagued by violent and racist images

https://www.theguardian.com/us-news/2025/oct/04/openai-sora-violence-racism
1•andsoitis•4m ago•0 comments

Carve Up a Mr Robot FSociety Mask Pumpkin for Halloween [Template][pdf]

https://decider.com/wp-content/uploads/2015/10/print-fsociety.pdf
1•sans_souse•6m ago•0 comments

How One AI Model Creates a Physical Intuition of Its Environment

https://www.quantamagazine.org/how-one-ai-model-creates-a-physical-intuition-of-its-environment-2...
1•pykello•6m ago•0 comments

A Year After 'Pokémon Go,' Where Are the Augmented-Reality Hits? (2017)

https://www.wsj.com/articles/a-year-after-pokemon-go-where-are-the-augmented-reality-hits-1499284437
2•sogen•16m ago•0 comments

Space Mission Options for Reconnaissance and Mitigation of Asteroid 2024 YR4

https://arxiv.org/abs/2509.12351
2•bookofjoe•19m ago•0 comments

What Caused Flooding in Dubai

https://www.bloomberg.com/features/2025-dubai-storm-floods-cloud-seeding-geoengineering/
2•tpolm•21m ago•0 comments

Probiotics Finder

https://www.probioticfinder.org/
3•cubefox•27m ago•0 comments

Mic-E-Mouse, Covert Eavesdropping Through Computer Mice

https://sites.google.com/view/mic-e-mouse
2•Timothee•29m ago•0 comments

Unchecked and Unaccountable: How DOGE Jeopardizes Americans' Data

https://www.hsgac.senate.gov/media/dems/peters-report-finds-that-doge-continues-to-operate-unchec...
4•defrost•31m ago•0 comments

AI-powered open-source code laundering

https://github.com/SudoMaker/rEFui/blob/main/HALL_OF_SHAME.md
2•genkiuncle•36m ago•0 comments

The only AI curve that matters

https://www.exponentialview.co/p/the-only-ai-curve-that-matters
1•jedharris•36m ago•2 comments

Greater Manchester Police home working ban amid 'key jamming' probe

https://www.bbc.com/news/articles/c4gw0wyjxxno
5•jnord•58m ago•0 comments

Borehole Oscillators

https://www.gregegan.net/SCIENCE/Borehole/Borehole.html
4•sohkamyung•59m ago•0 comments

WInfast – Upwork Copilot

https://www.winfast.app/
2•myonim•1h ago•1 comments

Countries with highest levels of hearing loss have lowest use of hearing aids

https://bmjgroup.com/countries-with-highest-reported-levels-of-hearing-loss-have-lowest-use-of-he...
3•geox•1h ago•0 comments

My DjangoCon US 2025

https://www.paulox.net/2025/10/05/my-djangocon-us-2025/
2•pauloxnet•1h ago•0 comments

Whiteboarding with AI

https://jrfernandez.com/whiteboarding-with-ai/
2•dirtyhand•1h ago•0 comments

Why Microservices Are a Symptom, Not a Cure – and What the Future Looks Like

https://axx83.substack.com/p/a-manifesto-for-structural-oriented
5•axx83•1h ago•0 comments

Russia's Hybrid War Against NATO Ramping Up: Danish Intelligence

https://www.twz.com/sea/russias-hybrid-war-against-nato-ramping-up-danish-intelligence
9•breve•1h ago•1 comments

An Alternative to Cryonics:An Introduction to Sparks Brain Preservation

https://biostasis.substack.com/p/an-alternative-to-cryonics
3•paulpauper•1h ago•0 comments

Chips Paradox: Potatoes good, Oilz good, Chips bad

https://theheartattackdiet.substack.com/p/chips-paradox
4•paulpauper•1h ago•0 comments

Wirehead: Personalized AI for generating images without prompts

https://wirehead.org
1•crainstorm•1h ago•0 comments

Rubis

https://rubis.onrender.com/login
2•lynx_reseau•1h ago•0 comments

How to Train an LLM to Do Proofs: Beyond Verifiable Rewards

https://tobysimonds.com/research/2025/09/29/Proofs.html
2•tamassimond•1h ago•0 comments

How do drivers react to partisan bumper stickers?

https://www.frontiersin.org/articles/10.3389/fpos.2025.1617785
19•PaulHoule•1h ago•1 comments

NSA and IETF: Can an attacker purchase standardization of weakened cryptography?

https://blog.cr.yp.to/20251004-weakened.html
6•zdw•1h ago•1 comments