Abliteration.ai is making a business out of removing AI guardrails

- Abliteration.ai is launching a business focused on making AI models without safety guardrails more accessible to the public.
- The service utilizes "abliteration," a long-standing open-source technique used to remove refusal mechanisms from open-weight models, many of which are already hosted on Hugging Face.
- The company argues that providing these unrestricted tools to defenders is essential for improving cybersecurity by mirroring the capabilities of bad actors.
- This move highlights an ongoing tension in the AI industry between maintaining safety guardrails and promoting open-access research for security purposes.
Sources & Citations
1 sourceMore Stories
Venture Capital & Startup Funding Roundup, September 3, 2026: Base10 Partners, Gradient Ventures, Point72, Y Combinator & More - Tech Startups
• On September 3, 2026, Boston-based healthtech startup Elucid raised $55 million in a Series D round led by IAG Capital Partners and Elevage Medical Technologies, bringing its total funding to approximately $185 million. • Tel Aviv-based cybersecurity firm Guardio (formerly Castle) secured $40 million in new funding, pushing its post-money valuation to roughly $1.1 billion.
Read original · techstartups.comSui Launches $10M Fund For AI And DeFi Builders
• The Sui Foundation has launched a $10 million ecosystem fund specifically targeting developers building decentralized AI infrastructure and DeFi-native protocols. • The initiative aims to attract more builders to the Sui ecosystem by leveraging its high-throughput blockchain and Move-based programming language.
Read original · bitcoinist.comOpen Source AI News
• September 2026 reports on open source AI highlight how founders can utilize portable open models to reduce operational costs, enhance data protection, and increase workflow control. • Key security recommendations include implementing default permission limits, isolating test environments, rotating credentials, and requiring human approval for external tool actions.
Read original · blog.mean.ceo
Mean CEO's BLOGAsia’s economies avert energy shock via fiscal buffers, AI boom | Ruchi Bhatia & Srinidhi Ragavendran
• Economies across Asia, including India, Malaysia, and Australia, posted solid growth last quarter, defying worst-case economic scenarios. • Growth was sustained by government interventions to secure energy supplies and the use of fiscal buffers to protect households from energy shocks.
Read original · businessmirror.com.phChinese AI firm Moonshot files confidentially for Hong Kong ipo targeting $3 billion
• Beijing-based AI firm Moonshot, the developer of the Kimi large language model, has confidentially filed for an initial public offering (IPO) in Hong Kong. • The company is targeting to raise approximately $3 billion through the prospective float.
Read original · dimsumdaily.hkHidden China risks are emerging in America’s multibillion-dollar AI data center boom
• U.S. hyperscalers are investing billions of dollars to rapidly expand AI data center infrastructure to meet growing demand. • These facilities rely heavily on Chinese-linked supply chains for critical hardware, including batteries, transformers, and optical gear.
Read original · cnbc.comSingapore captures 82.5% of regional blockchain funding
• Singapore has secured 82.5% of all-time blockchain equity funding across Southeast Asia, according to data from the private-market platform Tracxn. • The city-state's overwhelming dominance in regional investment highlights its position as the primary hub for blockchain innovation and capital.
Read original · sbr.com.sg
Singapore Business ReviewTechCrunch Disrupt: What’s next for AI and software development
• TechCrunch Disrupt 2026 will take place from October 13-15 in San Francisco, focusing on the evolution of AI and software engineering. • The event features over 200 conference sessions across six stages, led by more than 250 technology leaders.
Read original · infoworld.com
InfoWorldUS government backs OpenAI in New York Times copyright case By Reuters
• The US government filed a brief on Tuesday supporting OpenAI in its ongoing copyright lawsuit brought by The New York Times. • Federal officials argued that training Large Language Models (LLMs) is "extraordinarily" transformative and falls under the fair use doctrine.
Read original · investing.comLLM Daily: September 03, 2026 • Buttondown
• OpenAI is preparing to launch "Astra," a new model utilizing a "recurrent depth" technique that allows it to reason beyond standard sequential thinking patterns. • AI safety experts have raised alarms, warning that this novel architecture may be significantly harder to audit and constrain than previous models.
Read original · buttondown.com$15M Mission to Alpha Centauri, Sam Altman's 4-Month AGI, The First Fully AI-Designed Chip
• The Fermi Explorer mission, a $15 million project to Alpha Centauri, was announced in an episode published September 2, 2026. • The mission is spearheaded by PSI, which simultaneously announced a $58 million seed funding round led by Breakthrough Energy Ventures.
Read original · finance.biggo.com