Meta’s HyperNova AR Glasses

After drawing complaints on its latest flagship model, CEO Sam Altman conceded that GPT-5 rollout was “a bit more bumpy than we’d hoped.” Now, the corporate is pushing a personality update meant to make GPT-5 sound warmer without crossing into fake friendliness. As a substitute of empty flattery, GPT-5 now drops subtle cues like “Good query” or “Great start”. CEO Sam Altman says deeper customization will follow, letting users tune ChatGPT’s style directly.

As OpenAI fine-tunes its products, Altman can be blunt about market risks, comparing today’s AI hype cycle to the dot-com bubble, saying investors are “overexcited a few kernel of truth.” Heavyweights from Alibaba’s Joe Tsai to Bridgewater’s Ray Dalio have voiced similar warnings prior, and Apollo economist Torsten Slok argues valuations in today’s S&P 500 tech giants may already exceed Nineties excess.

Despite those warnings, capital continues to flood in. OpenAI employees are preparing to sell roughly $6 billion in shares to SoftBank, Thrive Capital, and Dragoneer Investment Group, in a deal that might raise the corporate’s valuation to $500 billion, up sharply from $300 billion just months ago. Read more.

Meta’s next AR gamble is nearly here. Leaked details on its “Celeste” smart glasses point to a right-lens HUD, built-in camera, and a neural wristband (“Ceres”) that picks up muscle signals for gesture controls. Early chatter pegged the worth north of $1,300 but Bloomberg now says closer to $800, with the wristband bundled, to avoid one other Quest Pro-style flop.

That undercuts earlier speculation while keeping it higher than most consumer smart glasses ($269–$649), though the package guarantees more capability: app support much like Quest 3, touch and hand-gesture navigation, and a built-in camera. Prescription lenses and styling options will raise the worth further. With Meta Connect set for September 17, the Celeste glasses are expected to debut there, with preorders likely before October shipping. Pricing will determine if Celeste is seen as a reasonable on-ramp to AR or one other overreach in Meta’s long bet on wearables. Read more.

OpenAI’s first open-weights release in six years is already being bent in unexpected directions. Lower than two weeks after OpenAI launched its Apache 2.0–licensed gpt-oss family, researcher Jack Morris (Cornell Tech/Meta) released gpt-oss-20b-base a stripped-down version of the 20B model that removes OpenAI’s “reasoning alignment” and restores raw, pretrained behavior. As a substitute of stepping through chain-of-thought logic, the bottom model simply predicts the following token, yielding faster, freer, less filtered text, including responses the aligned model would normally block.

Morris achieved this by applying a LoRA update to simply 0.3% of the network’s weights (three MLP layers, rank 16), trained on 20,000 FineWeb docs over 4 days on eight NVIDIA H200 GPUs. It’s now continue to exist Hugging Face under an MIT license, open for anyone to check or commercialize. Researchers get a clearer view of how LLMs behave before alignment but there’s a tradeoff: more unsafe, uncensored, and copyright-spilling behavior. Nevertheless, this shows how briskly open-weight models may be remixed, and the way little compute it takes to peel back alignment. Read more.

Anthropic just gave its latest Claude models a approach to walk away from conversations. Claude Opus 4 and 4.1 can now terminate chats outright, not for on a regular basis disagreements, but in “rare, extreme” cases where users keep pushing for things like child sexual abuse material or step-by-step guides to mass violence. Anthropic says this isn’t about shielding people, but about protecting the model itself, coming out from its “AI welfare” research program, which explores whether systems needs to be allowed to exit harmful interactions.

Anthropic is obvious it doesn’t think Claude is sentient, but says it’s testing low-cost guardrails in case future AI moral status isn’t hypothetical. When Claude does cut off a thread, you may still open a brand new chat or revise prompts. The block is scoped to that one conversation. It stays an experiment for now, and Anthropic is actively gathering feedback. Read more.

Colin W: We’re a small women’s fashion brand that has recently moved from a wholesale model to an e-commerce model. Do you could have any AI tools that you just would recommend that might help reduce the workload to run the business, especially in content creation, videos and stills, but definitely in anything that might help reduce overhead costs, eg labour?

Anonymous: What’s probably the most effective approach to minimize the error rate when using an LLM (GenAI bot) – i.e. Using a RAG + Anchoring (referencing) to an actual webpage, other? Also, is there a model / prompting / other that has proved to have 0% error rate?

There’s no approach to hit 0% error with an LLM, but probably the most effective setup is a decent RAG pipeline with verification and abstention: retrieve from a high-quality corpus with hybrid search, make the model quote sources directly (and refuse if not found), use constrained decoding and tools for math/code, then add a verification loop or cross-model check. Mix that with confidence thresholds so the system can abstain as a substitute of hallucinate. This drives error rates very low, however the only true “zero” comes from returning exact source text, or prompting it to return uncertain cases to a human.

Reply to this email to have your query featured within the newsletter!