In partnership with
News of the day
1. Opus 4.8 calls itself modest but hides Mythos, the extra level the OECD feared. And it now sells honesty. → Read more
2. NBA to implement AI for automatic out-of-bounds calls, inspired by Hawk-Eye, to improve accuracy and reduce stoppages. → Read more
3. Meta rolls out global subscriptions for Instagram, Facebook, and WhatsApp, offering premium features and testing AI and business plans. → Read more
4. Mistral AI targets 1 gigawatt computing power by 2030, signaling massive energy needs for AI development and infrastructure growth. → Read more
Our take
Hi Dotikers!
Yesterday, the OECD was explaining to us, charts in hand, that white-collar workers were still safe, with one condition attached: that the models do not gain one extra cognitive level. Twenty-four hours later, Anthropic releases Claude Opus 4.8 and casually ticks the very box the report was dreading.
On paper, the update wants to look low-key. Anthropic says so itself: a modest but tangible improvement over Opus 4.7, at the same price, with a few benchmark gains in coding, reasoning and agentic tasks. Thrown in along the way are an effort control on the user's side, dynamic workflows in Claude Code able to launch hundreds of subagents in parallel, and a fast mode three times cheaper than before. A routine version bump, on the face of it.
Except the real selling point lies elsewhere, and it is almost cheeky: Anthropic is marketing honesty as a feature. Opus 4.8 is said to be four times less likely than its predecessor to let a flaw slip through in the code it has just written, and to admit more readily that it does not know. In a market where GPT-5.5 and the rest still hallucinate with gusto, turning caution into a commercial asset is hardly absurd. It is even clever enough to make you wonder why nobody had seriously tried it before.
The detail that truly matters is tucked away at the bottom of the announcement. Anthropic is preparing a class of models more intelligent than Opus, named Mythos, already being tested by a handful of organizations on cybersecurity matters, and held back long enough to beef up the guardrails. In other words, that famous extra level the OECD filed under hypothetical future already exists, in a drawer. Yesterday's wave is now only waiting for a safety green light.
Small visible steps, big quiet leaps: the company has clearly found its rhythm.
Alex.
Headline: Your marketing stack reports to one place now.
Your media buyer opens Slack at 8am. There's already a cross-platform brief in #growth: Google Ads spend vs. ROAS, Meta CPA by campaign, Stripe revenue by channel. Viktor posted it at 6am. Nobody asked for it.
Same colleague caught a spend spike overnight on your brand campaign. Flagged it before anyone logged in. The problem was handled before the first standup.
Your strategist reviews trends. Your account manager checks attribution. Same Slack channel. Same colleague. Before anyone's first coffee.
Google Ads, Meta, Stripe. One message. No Looker. No Data Studio. No dashboard tab left open since Tuesday.
11,000+ teams use Viktor daily. SOC 2 certified. Your data never trains models.
Meme of the day
