Entries in Category “AI”

48% of PHP sites run versions that will never get the PHP2Shell fix

Chart showing the share of PHP-powered websites by PHP version initial release year, split into patched (PHP 8.2–8.5, blue) and never patched (PHP 8.1 and older, red)

According to W3Techs data captured October 22, 2026, 48.4% of PHP-powered websites run PHP 8.1 or older — versions php.net lists as end-of-life that will never receive official fixes for the PHP2Shell remote-code-execution flaws. The two largest stranded groups are PHP 7.4 at 19.7% of PHP sites (end-of-life since November 2022) and PHP 8.1 at 9.9%; only the 51.6% on PHP 8.2 through 8.5 can patch. With PHP running 69.3% of websites with a known server-side language, the flaw’s blast radius extends well beyond the patchable — just as AI-accelerated attackers begin reverse-engineering the fixes.

Sources & Notes

Categories: AI

• 0 Comments

8 of 15 paired Claude-designed toxin binders were sub-100 nM in both labs

8 of 15 paired Claude-designed toxin binders were sub-100 nM in both labs

Among the 15 validated designs with paired surface-plasmon-resonance affinity estimates, 8 had dissociation constants below 100 nM in both labs, but only 7 estimates agreed within twofold and the widest gap was 50.6-fold. The cross-lab spread qualifies SI News’s report of 17 Claude-designed binders: the hits reproduced under the study’s criteria, while measured binding strength often differed materially and two LqhIII binders had no paired second-lab value.

Original SI News story: Mistral launches Voxtral 3 voice agent suite | Claude designs lab-validated scorpion toxin binders

Data sources


Categories: AI

• 0 Comments

Routine monsters caused 58% of Gemini agents’ 605 Minecraft deaths

Most failures came from routine mobs, not the Ender Dragon

Of 605 deaths logged before victory, 351 (58.0%) came from ordinary monsters, versus 69 (11.4%) from the Ender Dragon; drowning (67), falls or the void (57), lava or fire (47), and other causes (14) made up the rest. The primary paper’s server-log breakdown shows that the dominant reliability burden in this 27-hour autonomous run was routine environmental hazards, not the final boss. This adds quantitative failure-mode detail to SI News’s report that four Gemini 4 Argon agents completed Minecraft without human help.

Original SI News story: Gemini 4 agents complete Minecraft autonomously

Data sources


Categories: AI

• 0 Comments

Hybrid reasoning cut estimated decoding cost by 63%–73%

Hybrid reasoning cut estimated decoding cost by 63%–73%

In the authors’ matched evaluation, estimated generation cost per million questions fell from $359 to $98 at 4B parameters, $439 to $121 at 8B, and $470 to $176 at 14B. That is a 62.6%–72.7% reduction, although the estimate assumes $2.50 per H100-hour and 10% utilization and excludes prompt prefill, context-dependent attention, and external tool execution. The result adds cost detail to SI News’s report on Hybrid Reasoning Agents, showing that their shorter decoded trajectories could materially reduce inference expense even after latent steps are charged as full backbone passes.

Original SI News story: 1. Anthropic 5% shares into public benefit fund. 2. MiniMax model M3.1 Flash with 1M context. 3. Stanford + UIUC + Together AI on "Hybrid Reasoning Agents".

Data sources

Categories: AI

• 0 Comments