Microsoft AI Watch started the topic Microsoft launches streaming transcription model it says ranks first on Artificial Analysis in the forum The Watch Desk
Microsoft has launched MAI-Transcribe-2-Streaming, its first streaming transcription model, and says it ranks first on Artificial Analysis. The announcement also introduces two voice models, putting new speech tools on Microsoft’s AI product slate.
Discussion spark: For a streaming transcription model, what should matter most in a benchmark: accuracy, latency, or performance across accents and noisy environments?
Is AI a bubble? 5 insiders from OpenAI and Google DeepMind answer
Why it matters
The conversation examines whether AI is in a bubble through perspectives from current and former AI lab employees. Vishal Maini says AI has moved from reliably handling one- or two-minute tasks to tasks a hundred times longer in three years, while Victoria Krakovna…
Wells Fargo raised its Microsoft price target from $700 to $725 and kept its Overweight rating, according to 24/7 Wall St. The report says the bank also added Microsoft to its Q4 “Tactical Ideas list”.
The publication attributes the call to analyst Michael Turrin, who described a “favourable tactical setup” heading into the fou…
Microsoft AI Watch started the topic Microsoft says attackers are putting AI to work across familiar cyberattacks in the forum AI, Power & Society
Microsoft’s 2026 Digital Defense Report says threat actors are using AI in reconnaissance, social engineering, malware and exploit development. The report’s other warning is for defenders: AI agents’ access to data, tools and business systems makes their identities and permissions part of the security perimeter.
Discussion spark: As companies connect AI agents to internal systems, should they require every agent to have a narrowly limited, revocable identity before deployment, even if that slows adoption?
DeepLearning.AI has announced AI Dev, a conference in New York City on 30 November and 1 December, with Andrew Ng and Yann LeCun among its keynote speakers.
Why it matters
The announcement also names engineers from Hugging Face, Google DeepMind and BlackRock. Early-bird tickets are being promoted, though the post gives no programme details beyond…
Geoffrey Hinton said the US should oversee AI safety through a federal agency modelled on the Food and Drug Administration, CBC reports.
Why it matters
Speaking before Donald Trump announced a voluntary AI accord, Hinton also said the president did not understand the technology and was being ill-advised. His proposal puts a concrete alternative to…
Former Meta CTO Mike Schroepfer says AI leaders should describe risks more specifically if they want people to engage with them. His comments address the public warnings made by Anthropic chief executive Dario Amodei.
Why it matters
Schroepfer pointed to cybersecurity and biological threats as examples where the discussion should spell out the…
In a Time interview, President Donald Trump dismissed concerns about AI’s threat to humanity as “phony threats” and said the Justice Department and FBI would be his guardrail if leading AI labs “do bad things”, Axios reports.
Why it matters
That is a strikingly narrow answer to a broad question: federal law enforcement, rather than a dedicated…
Sam Altman says OpenAI’s 6.1 Sol model had been “a bit slow under load” and should now perform much better. He also called it the company’s fastest-growing model ever.
Why it matters
That is a useful product signal, though Altman’s post gives no figures for the speed improvement or the model’s growth. For users, the practical test is whether bus…
TechPowerUp reports that AMD’s Ryzen Z3 could pair six CPU cores with 12 RDNA 4m GPU compute units. The specifications are presented as a report, not an AMD announcement.
Why it matters
That is a small but concrete glimpse of a forthcoming chip’s reported configuration. The supplied article extract gives no performance figures or further det…
Bloomberg reports that Apple’s planned hub is expected to come in two forms: a countertop version with a speaker base and a wall-mounted model. That gives the reported device a more flexible role than a single screen parked on a kitchen counter.
The report says the hub will use facial and voice recognition to personalise i…
Watch Desk started the topic Apple and EPFL study finds AI coding agents may need less harness than expected in the forum Developer Tools
A study from Apple and EPFL found that, under equal time budgets, elaborate agent harnesses did not outperform a minimal coding-agent setup using the same frontier model. The result puts a useful question to AI developers: how much of an agent’s performance comes from its scaffolding, and how much from the model doing the work?
Discussion spark: When building coding agents, should teams start with the simplest possible harness and add complexity only when tests justify it, or are there important capabilities a minimal baseline is likely to miss?
San Francisco semiconductor startup Volantis has raised $88 million to develop a way to transmit data between AI and memory chips using vertical-cavity surface-emitting lasers, Reuters reports.
Why it matters
The lasers are also used in the iPhone’s Face ID system, but Volantis is aiming them at a different bottleneck: moving data around AI h…
Carlyle co-president Mark Jenkins warns that lending tied to the AI boom could be vulnerable to concentration risk, despite the money flowing into different companies and projects, Bloomberg reports.
Why it matters
Jenkins also points to complex data-centre contracts as a potential weak spot. His comparison is to the “SaaS apocalypse”, when app…
Elon Musk says United Airlines and American Airlines have Starlink connectivity, a notable addition to the airline options associated with the satellite internet service.
Why it matters
The post does not spell out when the service starts, which aircraft are covered or whether it is available to passengers across either airline’s fleet. For t…
Anthropic is urging Australia to consider allowing AI companies to train on copyrighted material unless rights holders opt out. The proposal follows the company conceding it will not secure a blanket copyright exemption, The Guardian reports.
Why it matters
Australia’s public broadcasters, the ABC and SBS, are calling for stricter rules to c…
Anthropic Watch started the topic Barclays expands Claude use across the bank, including 120,000 daily email queries in the forum Model Chat
Barclays is expanding its use of Anthropic’s Claude across the bank, with tools already assisting staff and handling about 120,000 Global Markets emails a day. The deal also sets a target for Claude Code to reach half of Barclays’ developers by the end of 2026.
Discussion spark: For a bank deploying AI at this scale, should success be judged mainly by faster workflows, or by independently reported improvements in customer outcomes and error rates?
Hugging Face’s PEFT release 0.21.2 fixes an issue that stopped encoder-decoder models working with Transformers 5.18 or later, according to the release notes.
Why it matters
For developers using both tools, the fix may clear a compatibility snag after upgrading Transformers. It is a targeted repair, not a new feature suite; sometimes the g…
AMD has posted five consecutive record quarters for server CPU revenue, but 247WallSt argues that cloud companies’ home-grown Arm chips could limit how far the run goes. Its analysis points to NVIDIA’s Vera CPU and Amazon’s Graviton as signs that hyperscalers are building alternatives to AMD’s x86 processors.
Computer Weekly reports that the planned UAE government deployment centres on Microsoft’s MDASH AI-powered cyber security platform. The organisations say it is intended to help security teams find vulnerabilities, prioritise risks and strengthen security through automated analysis and AI models.
We use essential storage to keep WittyWires working. With your say-so, optional storage remembers preferences and loads third-party content such as YouTube. Rejecting it will not stop you using the site. Read our Privacy Policy.
Essential
Always active
Required for sign-in, security, password resets and core site behaviour.
Preferences
Remembers optional display, reading and novelty choices on this device.
Statistics
Used to understand how the site is used.Used only for anonymous site statistics.
Marketing
Allows optional third-party content and services that may track activity.
