diff --git a/README.md b/README.md
index eb563ce..817db67 100644
--- a/README.md
+++ b/README.md
@@ -198,6 +198,7 @@ This should start a basic webserver that would allow us to view the game state a
 
 
 THINKING ABOUT::
+* fix time, dont need to pass time everywhere!
 * finalize on logging, data vs flat
 * remove the llm calls in init
 * refactor add_agents bc of jumping
diff --git a/configs/defaults.json b/configs/defaults.json
index 2cf9faf..217740f 100644
--- a/configs/defaults.json
+++ b/configs/defaults.json
@@ -30,7 +30,7 @@
   "MEMORY_QUERY_COUNT": 20,
   "CONVERSATION_DEPTH": 5,
   "CONVERSATION_COOLDOWN": 5,
-  "LLM_IMPORTANCE": 0,
+  "LLM_IMPORTANCE": 1,
   "ALLOW_OBSERVANCE": 1,
   "DEFAULT_NAMES": [
     "Alice",
diff --git a/src/agents.py b/src/agents.py
index 2e7ea29..d934141 100644
--- a/src/agents.py
+++ b/src/agents.py
@@ -66,24 +66,28 @@ def update_goals(self):
         # this assumes 8+ importance is always worth changing /reevaluating goals
         # do i need to update my goals. if so, give me new goals
         relevant_memories = self.getMemories(None,unix_to_strftime(self.matrix.unix_time))
-        relevant_memories_string = "\n".join(f"Memory {i + 1}:\n{memory}" for i, memory in enumerate(relevant_memories)) if relevant_memories else ""
+        relevant_memories_string = "\n".join(f"Memory {i + 1}: \"{memory}\"" for i, memory in enumerate(relevant_memories)) if relevant_memories else ""
         prompt = f'''
-About {self}:
+you are a character in a world.
 {self.getSelfContext()}
 
-{self}'s goal is {self.goal}.
+goal:"{self.goal}".
 And {self}'s recent memories:
 {relevant_memories_string}
-Write out what my new overarching goal should be in a short sentence.
+Write out what my new goal should be in a short sentence given recent memories.
+Take into account what happened recently to help you focus.
 Write in the first person from the point of view of {self}.
-Write in natural short response style.
+Write out the goal only in json:
+{{"goal":"goal"}}
     '''
         msg = llm.generate(prompt)
         self.addMemory("observation",f"I updated my goal to be \"{msg}\"", unix_to_strftime(self.matrix.unix_time), random.randint(5, 8))
         self.goal = msg
 
+    def decide(self):
+        self.matrix.llm_action(self,self.matrix.unix_time)
+
     def ask_meta_questions(self, timestamp):
-        #relevant_memories = self.memory[-50:]
         relevant_memories = self.getMemories(None, timestamp)
         relevant_memories_string = "\n".join(f"Memory {i + 1}:\n{memory}" for i, memory in enumerate(relevant_memories)) if relevant_memories else ""
         vars = {"agent":self,"relevant_memories_string":relevant_memories_string}
@@ -93,7 +97,6 @@ def ask_meta_questions(self, timestamp):
         self.meta_questions.extend(x[1] for x in m if x[1] not in self.meta_questions)
 
     def evaluate_progress(self):
-        #relevant_memories = self.memory[-50:]
         relevant_memories = self.getMemories(self.goal, unix_to_strftime(self.matrix.unix_time))
         relevant_memories_string = "\n".join(f"Memory {i + 1}:\n{memory}" for i, memory in enumerate(relevant_memories)) if relevant_memories else ""
         prompt = f'''
@@ -121,7 +124,8 @@ def evaluate_progress(self):
         if score and explanation:
             self.addMemory("meta", explanation, unix_to_strftime(self.matrix.unix_time) , 10)
             if score and int(score) < 3:
-                self.meta_cognize(unix_to_strftime(self.matrix.unix_time), True)
+                #self.meta_cognize(unix_to_strftime(self.matrix.unix_time), True)
+                pass
 
     def meta_cognize(self,timestamp,force=False):
         #if not force and random.randint(0, 100) < 50:
@@ -549,8 +553,8 @@ def perceive(self, other_agents, environment, timestamp):
     def addMemory(self, kind, content, timestamp=None, score=None,embedding=None):
         memory = None
         if timestamp is None:
-            timestamp = datetime.now()
-            timestamp = timestamp.strftime("%A %B %d, %Y")
+            timestamp = datetime.utcfromtimestamp(self.matrix.unix_time).strftime("%A %B %d, %Y")
+
 
         if kind == "observation":
             if (self.matrix is not None and self.matrix.allow_observance_flag == 1):
@@ -669,6 +673,9 @@ def __init__(self, dictionary):
             min_relevancy_score = min(min_relevancy_score, relevancy_score)
             max_relevancy_score = max(max_relevancy_score, relevancy_score)
 
+            #print("PPPP")
+            #print(mem.content)
+            #print(mem.kind)
             recency_score = Memory.calculateRecencyScore(mem.last_accessed_at, timestamp)
             min_recency_score = min(min_recency_score, recency_score)
             max_recency_score = max(max_recency_score, recency_score)
diff --git a/src/matrix.py b/src/matrix.py
index 579bf3f..8f90627 100644
--- a/src/matrix.py
+++ b/src/matrix.py
@@ -25,6 +25,7 @@
 def set_globals(config):
     for key, value in config.items():
         globals()[key] = value
+
 class Matrix:
     def __init__(self, config={}):
         set_globals(config)
@@ -49,6 +50,8 @@ def __init__(self, config={}):
         self.cur_step = 0
         self.current_substep = 0
         self.unix_time = DEFAULT_TIME
+        #print(f"PPPP {self.unix_time} {type(self.unix_time)}")
+        #print(f"PPPP {LLM_IMPORTANCE} OOOOOO")
         self.status = "init"
         self.conversation_counter = 0
         self.sim_start_time = None
@@ -62,6 +65,7 @@ def __init__(self, config={}):
         self.perception_range = PERCEPTION_RANGE
         self.allow_movement = ALLOW_MOVEMENT
         self.model = MODEL
+
         self.replay = None
         self.add_to_logs({"step_type":"matrix_init","data":config})
         self.agent_locks = { agent: threading.Lock() for agent in self.agents }
diff --git a/src/memory.py b/src/memory.py
index bca17d8..d779e26 100644
--- a/src/memory.py
+++ b/src/memory.py
@@ -12,13 +12,19 @@ def __init__(self, kind, content, created_at, last_accessed_at, score=None,embed
         self.kind = kind
         self.content = content
         self.created_at = created_at
-        self.last_accessed_at = last_accessed_at
-        if not isinstance(self.last_accessed_at, str):
+        if not isinstance(last_accessed_at, str):
             print("last_accessed_at should be string timestamp format")
+            self.last_accessed_at = unix_to_strftime(last_accessed_at)
+            print(type(self.last_accessed_at))
+        else:
+            self.last_accessed_at = last_accessed_at
+        #print(f"WTF {content} {type(created_at)} {type(last_accessed_at)}   ")
+        #print(f"{type(LLM_IMPORTANCE)} {LLM_IMPORTANCE}")
         if LLM_IMPORTANCE == 0:
             self.importance = score
         else:
             self.importance = Memory.calculateImportanceScore(self.content, memories)
+
         self.recency_score = 1
         self.relevancy_score = 1
         self.overall_score = 1
@@ -107,6 +113,8 @@ def calculateRelevanceScore(memory_embedding, context_embedding):
 
     def calculateRecencyScore(last_accessed_at, time, decay=0.99):
         try:
+            #print(f"laa {last_accessed_at} {type(last_accessed_at)}")
+            #print(f"time {time} {type(time)}")
             last_accessed_at = int(datetime.strptime(last_accessed_at, "%Y-%m-%d %H:%M:%S").timestamp())
             time = int(datetime.strptime(time, "%Y-%m-%d %H:%M:%S").timestamp())
             return float(decay ** int(time - last_accessed_at))
diff --git a/utils/utils.py b/utils/utils.py
index a277356..9e2a9e0 100644
--- a/utils/utils.py
+++ b/utils/utils.py
@@ -1,4 +1,5 @@
 from configs.configs import *
+#TODO remove this
 import requests
 import time
 import os
@@ -133,6 +134,8 @@ def generate(self, prompt, fallback="Llm Error"):
               "prompt": prompt,
               "stream": False
           }
+          #"temperature": 0.4,
+          #print(data)
           current_url = self.urls[self.call_counter % len(self.urls)]
           try:
               if self.model == "powerinf":
@@ -162,9 +165,9 @@ def generate(self, prompt, fallback="Llm Error"):
         self.log_calls(prompt, msg, end_time - start_time)
         self.call_times.append(end_time - start_time)
         if len(self.urls) > 1:
-          pd(f"current url {current_url}")
-        pd(f"INPUT:\n {prompt}")
-        pd(f"OUTPUT:\n {msg}")
+            pd(f"current url {current_url}")
+        #print(f"INPUT:\n {prompt}")
+        #print(f"OUTPUT:\n {msg}")
         pd(f"runtime: {end_time - start_time}")
         return msg
 
@@ -247,7 +250,6 @@ def prompt(self, prompt_name, variables=None, fallback="Llm Error"):
         return output
 
 llm = Llm()
-
 '''
 REDIS
 '''