[et_pb_section fb_built=”1″ _builder_version=”4.23.1″ _module_preset=”default” global_colors_info=”{}”][et_pb_row _builder_version=”4.23.1″ _module_preset=”default” custom_margin=”-1px|auto||auto||” global_colors_info=”{}”][et_pb_column type=”4_4″ _builder_version=”4.23.1″ _module_preset=”default” global_colors_info=”{}”][et_pb_text _builder_version=”4.23.1″ _module_preset=”default” custom_padding=”0px|||||” hover_enabled=”0″ global_colors_info=”{}” sticky_enabled=”0″]<\/p>\n
When typing this question into ChatGPT <\/span><\/span>I was impressed by the response that suggests deep insight into the relationship between the shape, weight and strength of a manhole cover, the challenges to manufacture and manipulate it as well as worker safety concerns. Surely an AI needs a model of the world and true reasoning capability to arrive at this answer?<\/span> [\/et_pb_text][et_pb_image src=”http:\/\/denayer.com\/blog\/wp-content\/uploads\/2023\/12\/manhole_covers.jpg” title_text=”manhole_covers” _builder_version=”4.23.1″ _module_preset=”default” global_colors_info=”{}”][\/et_pb_image][\/et_pb_column][\/et_pb_row][\/et_pb_section]<\/p>\n","protected":false},"excerpt":{"rendered":" When typing this question into ChatGPT I was impressed by the response that suggests deep insight into the relationship between the shape, weight and strength of a manhole cover, the challenges to manufacture and manipulate it as well as worker safety concerns. Surely an AI needs a model of the world and true reasoning capability to arrive at this answer?<\/p>\n","protected":false},"author":1,"featured_media":2062,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_et_pb_use_builder":"on","_et_pb_old_content":"","_et_gb_content_width":"","footnotes":""},"categories":[34],"tags":[],"class_list":["post-2059","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t
<\/span>
<\/span>But is this real reasoning, or is there\u00a0less\u00a0than meets the eye?\u00a0\u00a0It turns out that, through an intricate dance of data retrieval and pattern recognition the large language model is constructing its answer from the countless internet pages in its dataset where this HR interview question is discussed in detail.\u00a0\u00a0Many examples of<\/span> LLM <\/span>\u2018intelligence\u2019 can similarly be traced back to the dataset and the response accuracy drops immensely when we<\/span> <\/span>
<\/span>\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0reduce its ability to search its dataset, for example by changing the terminology used to describe the problem or<\/span>
<\/span>\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0propose problems outside of the dataset<\/span>
<\/span>
<\/span>The same is true when trying to solve planning problems where the LLM will be \u2018approximately retrieving\u2019 a plan from the many plans it has seen in its gigantic dataset. This is not reasoning but actually still\u00a0very useful,\u00a0as long as we have a way to validate these generated \u2018candidate plans\u2019, either by human experts or reliable planning software.<\/span>
<\/span>
<\/span>We can create significant business value with this technology as long as we<\/span>
<\/span>\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0don\u2019t get carried away by the hype and<\/span>
<\/span>\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0understand that we are very much in the man + machine era<\/span>
<\/span>
<\/span>So why are manhole covers round? Ask your favourite LLM!<\/span><\/p>\n